http://www.diva-portal.org

This is the published version of a paper published in Progress in and Planetary Science.

Citation for the original published paper (version of record):

McCrea, I., Aikio, A., Alfonsi, L., Belova, E., Buchert, S. et al. (2015) The science case for the EISCAT_3D . Progress in Earth and Planetary Science, 2(1) http://dx.doi.org/10.1186/s40645-015-0051-8

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version: http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-110324 McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 DOI 10.1186/s40645-015-0051-8

REVIEW Open Access The science case for the EISCAT_3D radar Ian McCrea1*, Anita Aikio2, Lucilla Alfonsi3, Evgenia Belova4, Stephan Buchert5, Mark Clilverd6, Norbert Engler7, Björn Gustavsson8, Craig Heinselman9, Johan Kero4, Mike Kosch10,11, Hervé Lamy12, Thomas Leyser5, Yasunobu Ogawa13, Kjellmar Oksavik14, Asta Pellinen-Wannberg4,15, Frederic Pitout16,17, Markus Rapp18, Iwona Stanislawska19 and Juha Vierinen20

Abstract The EISCAT (European Incoherent SCATer) Scientific Association has provided versatile incoherent scatter (IS) radar facilities on the mainland of northern (the EISCAT UHF and VHF radar systems) and on (the electronically scanning radar ESR (EISCAT Svalbard Radar) for studies of the high-latitude ionised upper atmosphere (the ). The mainland were constructed about 30 years ago, based on technological solutions of that time. The science drivers of today, however, require a more flexible instrument, which allows measurements to be made from the troposphere to the topside ionosphere and gives the measured parameters in three dimensions, not just along a single radar beam. The possibility for continuous operation is also an essential feature. To facilitatefuture science work with a world-leading IS radar facility, planning of a new radar system started first with an EU-funded Design Study (2005–2009) and has continued with a follow-up EU FP7 EISCAT_3D Preparatory Phase project (2010–2014). The radar facility will be realised by using phased arrays, and a key aspect is the use of advanced software and data processing techniques. This type of software radar will act as a pathfinder for other facilities worldwide. The new radar facility will enable the EISCAT_3D science community to address new, significant science questions as well as to serve society, which is increasingly dependent on space-based technology and issues related to . The location of the radar within the auroral oval and at the edge of the stratospheric vortex is also ideal for studies of the long-term variability in the atmosphere and global change. This paper is a summary of the EISCAT_3D science case, which was prepared as part of the EU-funded Preparatory Phase project for the new facility. Three science working groups, drawn from the EISCAT user community, participated in preparing this document. In addition to these working group members, who are listed as authors, thanks are due to many others in the EISCAT scientific community for useful contributions, discussions, and support. Keywords: EISCAT; EISCAT_3D; Radar; Incoherent scatter; Atmospheric science; Space physics; Plasma physics; Solar system research; Space weather; Radar techniques

Review it a magnetic field which interacts with the internally Introduction to EISCAT_3D generated magnetism of the Earth and other planets. Why do we need EISCAT_3D? The science of solar-terrestrial physics (STP) is con- The interaction between the and the Earth is vital cerned with understanding all aspects of the relationship to every aspect of human existence. As well as providing between the Earth and the Sun. In particular, it seeks to us with heat and light, the Sun supplies the energy understand all the different ways that energy from the which powers the motion of the Earth’s atmosphere and Sun is deposited in the environments of the Earth and oceans, governing our weather and climate. In addition, other solar system bodies, the processes by which that the Sun produces the solar —a stream of energetic energy is converted from one form to another, and particles which permeates the solar system, carrying with the combined effects of all these processes on our environment. * Correspondence: [email protected] Because of this broad remit, STP overlaps with many 1 RAL Space, STFC Rutherford Appleton Laboratory, Harwell, Oxfordshire, UK other areas of science, including atmospheric physics, Full list of author information is available at the end of the article

© 2015 McCrea et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 2 of 63

solar physics, and plasma physics. It also has many prac- provide the kind of observations that have never been tical goals, including the better prediction and mitigation possible with the existing EISCAT radars. of space weather to safeguard space-based technology, improved modelling of the Earth’s ionised atmosphere What is EISCAT_3D? for communications and global positioning applications, EISCAT_3D is a new kind of international research radar and a deeper understanding of the contribution of nat- for studies of the Earth’s upper atmosphere and the region ural variability to long-term and short-term global of interplanetary space surrounding the Earth. The new change. In a world ever more dependent on space-based radar system, to be located in northern Scandinavia, will systems, and in which an improved understanding of comprise several large fields of antennas, known as phased our environment is at the top of the scientific and polit- arrays, some of which will have both transmitting and re- ical agenda, solar-terrestrial physics has become a key ceiving capabilities while others will be passive receivers. science area for the twenty-first century. An artist’s impression of an EISCAT_3D site is shown in Progress in STP is largely driven by observations. Be- Fig. 1. According to the current design, there will be at cause the solar-terrestrial system is continuously varying, least five radar sites, of which at least one, known as the and the interplay between the various processes is so “core site” will include a transmitter. Because the arrays complex, current theoretical models by themselves can are designed to be modular, it will be easily possible only bring us a certain distance in understanding how to enlarge the sites or to add additional transmitting the Earth’s environment responds to solar influences. capabilities, as funding allows; hence, the already sub- Hence, there is a major requirement for high-quality stantial capabilities of EISCAT_3D could be even fur- data from continuous, multi-point observations of all ther expanded, provided that the initial antennas had the key regions of the Sun-Earth system, from the solar the bandwidth required to exploit the additional photosphere and corona to the and magneto- transmitter capabilities. sphere, through the ionosphere and thermosphere, and EISCAT_3D is designed to use several different meas- down to the middle and lower atmosphere. This requires urement techniques which, although they have each a number of different instruments, including satellites, been used elsewhere, have never been combined to- radars, optical imagers, and ground-based magnetic gether in a single radar system: measurements, working in combination to observe dif- ferent aspects of the coupled system and provide the in- Volumetric imaging The design of EISCAT_3D allows put data for the development of sophisticated modelling large numbers of antennas to be combined together to and forecasting techniques. Science on this scale can make either a single radar beam or a number of simul- only be done by international collaboration, and, al- taneous beams, via a process known as beam forming. though some observing instruments are relatively cheap, In radars, the properties of the beam form- the development of leading-edge instruments, such as ing can be changed very rapidly, so that the direction of spacecraft and radars, in most cases requires multi- the radar beams can change every few milliseconds. national funding and the coordinated effort of a world- These capabilities, possible because there are no moving wide research community. parts, are a huge advantage compared to the traditional The EISCAT (European Incoherent SCATer) Scientific Association has played a key role in providing world- leading radar systems for use by the whole international research community over the past 30 years. As well as carrying out its own independent experiments, EISCAT has run extensive observing programmes to support a variety of spacecraft and collaborated closely with the other incoherent scatter radars around the world, particu- larly with those funded and operated by the US National Science Foundation, as part of a global observing programme based around the Geophysical World Days, including the International Polar Year of 2007–2008. To solve the existing questions related to the upper, partly ionised, atmosphere and its coupling to the lower atmosphere, as well as to the Sun and the solar wind, a new type of incoherent scatter radar must be con- ’ structed. This new radar system, EISCAT_3D, will ex- Fig. 1 An artist s impression showing the core site of the EISCAT_3D facility ( credit: FFAB:UK) ploit the twenty-first century phased array technology to McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 3 of 63

kinds of radar using large dishes. They allow the radar steering. Other novel techniques, including passive radar either to look in multiple directions simultaneously or to measurements and active control of the beam shape, for “paint the sky”, repeatedly scanning a single beam instance to control the shape of phase fronts in the near through a range of directions, to build up quasi- field, should also be possible with EISCAT_3D. simultaneous images of a wide area of the upper atmos- phere in three dimensions. While radar systems with a Continuous monitoring EISCAT_3D will allow remote single, slow-moving, beam can only show us what is continuous operations, limited only by power consump- happening in a single profile of the upper atmosphere at tion and data storage. This is important for monitoring any one time, volumetric imaging allows us to see geo- the state of the atmosphere, especially as a function of physical events in their full spatial context and to distin- solar variability, as well as capturing unexpected events guish between processes which vary spatially and those that appear suddenly and are hard to predict. EISCAT_3D which vary in time. will be an essential tool for validating short- and long- term models for predicting the state of the space environ- Aperture synthesis imaging Most radars suffer from ment as well as global change. As the population of the the limitation of not resolving any structure smaller than Earth grows and demands on natural resources increase, the width of their narrowest (transmitter or receiver) and as our dependence on space assets continues to grow, beam, typically of order a kilometre for ionospheric ra- the need for accurate forecasting increases. dars. In EISCAT_3D, this limitation is overcome by div- iding the core site into a number of sub-arrays, the EISCAT_3D key capabilities signals from which will be continuously cross-correlated The EISCAT_3D radar system will have the following in a technique similar to the very long baseline interfer- key capabilities, which represent a unique set given by a ometry approach used by radio astronomers, except that single facility, both in scientific and technological terms: the baselines here are only a few hundred metres. The result of inverting the cross-correlations is a brightness 1. Resolution of space-time ambiguity: EISCAT_3D will function showing the distribution of backscatter down have simultaneous multiple-beam capability. This to very small scales, less than the width of a normal re- will resolve outstanding issues of spatial-temporal ceiver beam, distinguishing between situations where the ambiguity (e.g. the dynamics of dusty plasmas in the beam is evenly filled with electrons and those where the mesopause region and rapidly moving auroral struc- scatter comes from a few discrete intense structures, the tures, tracking and meteors). study of which could produce exciting new science. 2. 3D volumetric capability: EISCAT_3D will have 3D volumetric imaging capability throughout its field of Multi-static imaging Several of the EISCAT_3D sites view. Such capability is important for studying the will be passive receivers, located at distances between 50 variability, coupling, and energy dissipation between and 250 km from the core site. Like the core site, each the solar wind, , and atmosphere (e.g. remote site will be capable of generating multiple simul- Joule heating and field-aligned currents) as this taneous beams, and this will enable the transmitter beam coupling is a function of altitude, latitude, and from the core site to be imaged simultaneously over a longitude. large range of altitudes from a variety of different ob- 3. Sub-beam width measurements: The spatial scale of serving directions. This will make it possible to con- micro-physical processes is much less than the struct continuous height profiles of parameters such as current EISCAT UHF radar beam half-width of 0.5° vector velocity and ionospheric current density, or to (e.g. NEIALs, small-scale and black , meteor look for anisotropic scattering mechanisms, in a manner head echoes, PMSE). EISCAT_3D will have the that cannot be achieved by conventional radars in a way needed capability to perform interferometry with compatible with typical geophysical coherence times. multiple baseline angles and lengths, a technique already proven by the EISCAT Svalbard Radar, Scanning, tracking, and adaptive experiments Be- which can be used to investigate these phenomena. cause EISCAT_3D is so flexible compared to traditional 4. Increased sensitivity and the resulting temporal ionospheric radars, it will allow several new operating resolution: With its improved sensitivity and modes including the capability to track moving objects temporal resolution, EISCAT_3D will be able to such as meteors and space debris or to respond intelli- reach down to the sub-second timescales that are gently to changing conditions, for instance by changing known to exist in auroral features from optical the parameters of a scanning experiment. While some measurements. As Appendix B of Lehtinen et al. existing radars are beginning to use these techniques, (2014) makes clear, EISCAT_3D will be able to they are often limited by factors such as inertia-bound measure auto-correlation functions at sub-second McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 4 of 63

time resolutions, under high-SNR conditions, at and solar wind, or cosmic rays which often originate far accuracies better than a few tens of percent errors in away in space. In addition, atmospheric gravity waves multi-parameter fits. Being able to measure multiple launched by the effects of geomagnetic storms in the plasma parameters at such high time resolutions upper atmosphere can propagate downwards to lower may provide a key to unraveling some of the cause/ altitudes (e.g. Rees 1989; Kelley 2009). Conversely, a effect relationships in auroral forms. great deal of energy reaches the middle and upper at- 5. Continuous monitoring of solar variability on mosphere from below—the periodic solar thermal en- terrestrial atmosphere and climate: EISCAT_3D will ergy input and the gravitational interaction of the Sun, allow remote continuous operations, limited only by Earth, and Moon raises atmospheric tides (Kato 1989; power consumption and data storage. This is Hagan et al. 2003), which can play an important role in important for monitoring the state of the the coupling of the atmospheric layers; planetary waves atmosphere, especially as a function of solar (Andrews et al. 1987; Arnold and Robinson 1998; Pancheva variability, and essential for capturing unexpected et al. 2008) and acoustic gravity waves, excited by lower events that appear suddenly and are hard to predict. atmosphere processes such as weather systems and 6. Model validation for space weather and global mountain lee waves, can also propagate upwards, pro- change: EISCAT_3D will be an essential tool for viding an important link to dynamical processes in the validating models of the upper atmosphere and upper atmosphere (Holton 1982; Lindzen 1984; Becker space environment, which can be expected to have and Schmidtz 2002). increasing societal relevance over the coming years. The global circulation of the middle atmosphere tends to produce polar vortices in the winter hemisphere Atmospheric physics and global change (Brewer 1949; Dobson 1956; Dunkerton and Delisi 1986; Background Schoeberl et al. 1992; Fisher et al. 1993), isolating the The extent to which the geospace environment, coupling polar air from that at lower latitudes—the reason, for ex- through upper atmospheric processes, can influence the ample, that the ozone holes form preferentially over the lower atmosphere (and vice versa) is a controversial poles. While downwelling air dominates the middle at- issue, with very significant implications. Any attempt to mosphere in the winter hemisphere, the summer hemi- understand the transport of energy through the whole sphere is dominated by upwelling (see Fig. 2). Adiabatic atmosphere has to take account of energy propagating cooling of the rising air in the polar summer mesosphere both downwards and upwards, as shown schematically produces the lowest temperatures found anywhere in in Fig. 2. Energy passes downward in the form of solar the Earth’s environment (Reid 1989; Lübken 1999; Singer radiation, energetic particles from the magnetosphere et al. 2003).

Fig. 2 Schematic of energy transfer processes linking the upper and lower atmosphere: GCR galactic cosmic rays, TSI total solar irradiance, SEP solar energetic particles (Gray et al. 2010) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 5 of 63

In addition to coupling by waves, tides, winds, solar regions of the atmosphere (Fritts and van Zandt 1993; radiation, energetic particles, and chemical composition Holton et al. 1995; Hocking 1996; Hamilton 1996, 1999; changes (which can be initiated from either direction), McLandress 1998; Meriwether and Gardner 2000). Much the upper and lower regions of the atmosphere are elec- of this wave energy occurs at mesoscales, in atmos- trically coupled via a current circuit (Roble and Tzur pheric gravity waves with periods of less than an 1986; Rycroft et al. 2000, 2002) which includes iono- hour, speeds of order 100–200 m/s, and wavelengths spheric currents, thunderstorm activity, and the “clear of hundreds of kilometres, chiefly arising from pro- air” electric field of the lower atmosphere. This electrical cesses in the troposphere (Hooke 1986). In addition, coupling is presently very poorly understood. The larger scale tides and Rossby waves, with periods of a coupled atmospheric system has a large number of feed- few hours, also arise from lower altitudes and propa- backs, whose importance is presently uncertain and needs gate upwards into the thermosphere (Rossby et al. to be much better established. It is believed that cosmic 1939; Lindzen 1979; Forbes 1982a, b; Chelton and rays may be able to modulate cloudiness (Svensmark Schlax 1996). 1998; Carslaw et al. 2002; Tinsley 2008), that the strength At mesospheric altitudes, upgoing gravity waves break of the polar vortex can be modulated by solar cycle effects into turbulence (Fritts and Luo 1995; Prusa et al. 1996), (Haigh 1996, 1999, 2003; Labitzke 1987, 2001, 2003, 2004, heating the middle atmosphere and depositing their en- 2005), and that this may relate to the unexplained correla- ergy and momentum flux into the background mean tions between solar wind dynamic pressure and middle at- flow. This turbulence fundamentally affects the global mosphere parameters such as stratospheric wind speed circulation of the atmosphere through the mechanism of and temperature (Tinsley and Heelis 1993; Arnold and gravity wave drag (Holton et al. 1995). These effects are Robinson 2001; Pudovkin 2004). highly variable but are most pronounced in the summer The behaviour of the lower atmosphere also modifies hemisphere. The dissipation of wave energy from below the spectrum of waves reaching the upper atmosphere in is thought to provide the majority of the momentum a way which has not been well described, and consider- flux to the mesosphere and lower thermosphere (Ebel ably more information about these feedback mechanisms 1984). Under the right conditions, this turbulence can is required. Processes leading to the formation and trans- also be associated with the presence of mesospheric thin port of ozone-depleting chemicals such as odd hydrogen layers (see “Solar-terrestrial effects on middle atmos- and odd nitrogen species (Crutzen 1970; Crutzen et al. phere chemistry”). 1975; Rusch et al. 1981; Solomon et al. 1981; Callis and In addition to waves propagating upwards, another Natarajan 1986a, b; Seppälä et al. 2006; Verronen et al. class of waves is created at ionospheric altitudes by geo- 2006) are of profound importance, since UV absorption by magnetic activity (e.g. Richmond 1978; Hunsucker ozone fundamentally affects the heat balance of the mid- 1982). Some of these waves are at mesoscales, as de- dle atmosphere, which in turn affects wave propagation. scribed above, while some occur as travelling iono- Understanding these factors is thus fundamental to our spheric disturbances (LSTIDs and MSTIDs) with understanding of atmospheric chemistry and dynamics. wavelengths of over a thousand kilometres, periods on Two of the main aims for EISCAT_3D are to under- the order of an hour or more, and propagation speeds of stand the ways in which natural variability in the upper up to 1000 m/s. These waves generally propagate equa- atmosphere, imposed by the solar-terrestrial system, can torward from high latitudes, where they are formed by influence the middle and lower atmosphere, and to im- the deposition of energy into the upper atmosphere via prove the predictive capabilities of atmospheric models Joule heating, particle heating, or mechanical forcing of by providing higher resolution observations to replace the neutral atmosphere via the Lorentz force (Brekke the current parameterised inputs. This kind of “system 1979). Using the Arecibo radar, Djuth et al. (1997, 2004) science” requires other instruments too, including dif- have demonstrated how incoherent scatter techniques, in- ferent types of radar, ground-based optics, and satel- cluding the use of plasma line measurements, can reveal lite instruments such as limb sounders. It also the continuum of thermospheric gravity wave activity. requires good coordination between experimentalists These techniques will be highly relevant to EISCAT_3D. and modellers. EISCAT_3D, by virtue of its extraor- The combination of this wave and tidal activity is repre- dinary versatility, will sit at the centre of this network sented schematically in Fig. 3. Understanding its cumula- of instruments and models, providing a key underpin- tive effects, and in particular where the energy and ning data set for this undertaking. momentum flux of atmospheric waves is deposited, is cru- cial to understanding the energy coupling of the atmos- Dynamical coupling in the atmosphere phere. Although the upper atmosphere has a global-scale Dynamical coupling by winds, waves, and tides is one of circulation pattern, this is affected by smaller scale waves the most important mechanisms connecting the different and turbulence (Lindzen 1981; Pfister et al. 1993), making McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 6 of 63

Fig. 3 Schematic showing the role of wave processes in energy coupling between the atmospheric layers and latitude regions (Sato et al. 2010) it essential to understand small-scale effects because of thermosphere (e.g. 70–130 km) would provide very valu- their global-scale atmospheric consequences. A major role able experimental constraints for models, of a kind which for EISCAT_3D will be to provide the high-resolution ob- it has not been possible to obtain so far. servations to describe the wave climatology over a wide Improved measurements of both large-scale and small- range of scale sizes and altitudes, in order to provide scale dynamics are also critically important because many much better inputs to atmospheric models. As Fig. 4 shows, wave activity, whether it arises from above or below, can have significant effects on the at- mospheric temperature, with variations of tens of de- grees at mesospheric altitudes displaying wave-like structure (Nakamura et al. 1993; Lübken et al. 2009). EISCAT_3D, because of its good height coverage, will be an ideal tool to study the propagation and dispersion of these waves and to quantify their horizontal structure. In particular, EISCAT_3D will be a superb complement to the German Middle Atmosphere Alomar Radar System (MAARSY) facility, which began operating at Andøya in 2010 (Latteck et al. 2010). While MAARSY can observe wave and tidal structures in the mesosphere, EISCAT_3D can extend these observations upward above the meso- Fig. 4 Temperature variability in the middle atmosphere caused by pause, where MAARSY is not designed to measure. This atmospheric gravity waves. The red lines represent temperature ability to probe the variation of 3D tidal amplitudes over a measurements inside mesospheric thin layers. The green lines represent all other temperature measurements (Lübken et al. 2009) broad range of heights in the mesosphere and lower McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 7 of 63

transport phenomena in the upper atmosphere are hard are short lived and mainly affect the chemistry of the to explain without understanding the wave dynamics. For lower mesosphere and upper stratosphere, the longer example, Stevens et al. (2003) found that space shuttle lived odd nitrogen species have been shown to descend plumes in the lower thermosphere were transported much to altitudes as low as 25 km in the southern hemisphere faster than expected from current models. More recent (Callis et al. 1996; Randall et al. 1998). This descent often work by Yue and Liu (2010) has indicated that such obser- occurs during the late winter and early spring, though the vations can be explained by the superposition of tides and effect does not occur every year, for reasons which are un- planetary waves, and EISCAT_3D would be capable of clear, but probably related to the variability of middle at- providing the data to substantiate this type of conclusion. mospheric dynamics (Callis and Lambeth 1998). Because odd nitrogen (in particular) can be trans- Solar-terrestrial effects on middle atmosphere chemistry ported over long distances, the effects of ozone depletion The ionised part of the mesosphere, known as the D re- are not confined to the poles. Modelling suggests that gion, is extremely important because of its complex ozone modulation can even occur in the tropics under chemistry, which has significant implications for struc- conditions of high geomagnetic activity (Dobbin et al. ture and dynamics of the middle atmosphere. The most 2006). It has been speculated that the atmospheric energetic precipitating particles, such as solar protons or changes which follow such ozone depletions might ac- energetic electrons with energies in the range from tens of count for the statistical relationships between surface kiloelectronvolts to tens of megaelectronvolts, penetrate temperature and geomagnetic activity, which appear to down to D region altitudes, creating significant chemical exist at high latitudes (e.g. Rozanov et al. 2005; Seppälä changes, including enhancements in odd nitrogen and et al. 2009). Such relationships are undoubtedly real and odd hydrogen species (Rusch et al. 1981; Solomon et al. can be reproduced in models, though their precise 1981; Seppälä et al. 2006; Verronen et al. 2006). These causes are not yet clear. It is notable, however, that these particles, known as solar energetic particles (SEPs), are relationships do not appear to be the same at all longi- usually seen in the first minutes following a tudes. For example, as shown in Fig. 6, higher geomag- (Cane et al. 1986; Reames 1990) and can reach strato- netic activity seems to be associated with positive spheric altitudes (Jackman et al. 2005a, b) or even to surface temperature anomalies in regions such as ground level. Even less energetic particles (protons with and the Antarctic Peninsula; however, there is also evi- energies of tens of megaelectronvolts or electrons with en- dence that higher geomagnetic activity is statistically as- ergies of hundreds of kiloelectronvolts) can reach strato- sociated with lower surface temperatures in Greenland spheric altitudes down to around 50 km. and over the Southern Ocean. Whatever the process is, It is now well accepted that energetic particle precipi- linking surface temperature to geomagnetic activity, it is tation during storms can significantly modulate strato- likely to be complicated and indirect. spheric ozone levels, which in turn changes the In addition to energetic particle precipitation, ablation atmospheric heat balance and the dynamical coupling of meteors changes the chemistry of the mesosphere by between the atmospheric layers (Andrews et al. 1987; adding metallic ions and aerosols in the form of meteor Austin et al. 1992; Shine 1986; Shindell et al. 1998). En- smoke. The Sodankylä ion and neutral chemical model, ergetic particle precipitation promotes the formation of which represents the state-of-the-art in describing D re- odd hydrogen (H, OH, and HO2) and odd nitrogen spe- gion chemistry (Verronen et al. 2002; Enell et al. 2005), cies (NO, NO2,NO3,N2O3, CINO3, and HNO3) (Rusch includes some 400 chemical reactions between 65 ions et al. 1981; Solomon et al. 1981). These species are im- and 15 neutral species. Many of these reactions have not portant because they play a catalytic role in the destruc- been extensively observed in practice, and the complex- tion of ozone (Brasseur and Solomon 1986), severely ity of D region chemistry can make the interpretation of depleting the stratospheric ozone layer over a period of observations very difficult. as little as a few days during active geomagnetic condi- Changes in dynamical coupling are likely to be an tions (Solomon et al. 1983; Reid et al. 1991; Jackman et al. important factor in explaining how the atmosphere 2000). The modulation of ozone is very noticeable after in- responds to those changes in atmospheric chemistry tense geomagnetic storms, such as the “Halloween super- which modulate ozone densities. The heat balance of storm” of October 2003 (Seppälä et al. 2004; Tsurutani the middle atmosphere, largely determined by ozone, is et al. 2006), and is likely to have important consequences likely to be a key factor in regulating which waves are able over broad regions of the middle and upper atmosphere to propagate from the troposphere and stratosphere into (see Fig. 5). the mesosphere and thermosphere, and which are Alternatively, ozone depletion can occur more slowly reflected or dissipated (Graf et al. 1998; Gabriel et al. as thermospheric species descend into the stratosphere 2007). This affects not only the upper atmosphere but also (e.g. Clilverd et al. 2006). While odd hydrogen species the structure and dynamics of the lower atmosphere, McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 8 of 63

Fig. 5 Variations in ozone density (upper panel) and nitric oxide density (lower panel) at 46 km altitude, as observed by the GOMOS instrument on ENVISAT, following the intense solar particle events (SPE) during the Halloween superstorm of 2003 (adapted from Seppälä et al. 2004) because there is evidence that reflection of wave energy in processes which can influence weather and climate, the the stratosphere may be a precursor of certain types of un- upper altitude ceiling of these models needs to be raised usual weather event (Baldwin and Dunkerton 2001). to upper atmospheric heights, so that the geospace pro- It is increasingly accepted that, in order to produce cesses capable of creating odd hydrogen and odd nitro- models which fully include all of those stratospheric gen species can be fully included (e.g. Roble 2000,

Fig. 6 Left: Surface air temperature changes in the northern winter hemisphere from model calculation including energetic electron precipitation and NOx production (Rozanov et al. 2005). Right: Difference between surface air temperature for the high Ap (geomagnetic activity index) minus low Ap years from 1957 to 2006 (adapted from Seppälä et al. 2009) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 9 of 63

Schmidt et al. 2006). EISCAT_3D will provide the experiments have the potential to give information on a underpinning observations, especially concerning the range of atmospheric properties and constituents that frequency and extent of energetic particle events, which are not accessible by natural excitations. A somewhat re- will make such model improvements possible, and EIS- lated topic in the literature is microwave discharge in CAT_3D data will allow the results of the modelling to the stratosphere, which has been suggested for the miti- be validated. gation of ozone depletion (Akmedzhanov et al. 1995). When strong scattering layers such as PMSE or The transmit antenna of EISCAT_3D needs to be de- PMWE are absent, the D region is often very difficult to signed appropriately to enable sufficient focusing for ex- observe with radars because of the low density of free periments on ionisation and breakdown in the upper electrons in this region. In this respect, the high power, atmosphere. The capability to focus the radar beam for large aperture, and low frequency of EISCAT_3D will Joule heating studies would also be useful to the devel- give it an advantage over any other type of modern inco- opment of solar power satellites, which would beam the herent scatter radar (ISR), providing a facility with con- energy back to Earth by microwaves (Leyser and Wong siderable synergy to the already existing European assets 2009). These powerful beams should suffer as small an for mesospheric research, such as rockets, optical instru- energy loss in the atmosphere as possible. Focusing ex- ments, and other types of radar. Since most current ISRs periments would give information on the atmospheric are not capable of observing much below 80 km, a key response to very high frequency beams. challenge for EISCAT_3D will be to probe the extent to which these energetic particles can reach significantly Dynamical and chemical coupling in the mesosphere lower altitudes, especially during very disturbed condi- The mesosphere (at altitudes from about 50 to 90 km) is tions (Jackman et al. 2005a, b). This is particularly true a very important area of the atmosphere, critical to the for energetic electron precipitation, which is hard to understanding of energy coupling, but often very diffi- measure accurately in any other way, and is an import- cult to observe with radars. The mesosphere has a very ant factor for global atmospheric models. While EIS- complicated chemistry, characterised by neutral and CAT_3D might not be able to make the full range of charged molecular species and cluster ions consisting of required measurements by itself, its data, in conjunction ice and meteoric dust. This chemistry is controlled with measurements from lidars and other types of radar, partly by energetic particle precipitation and also by dy- will provide unprecedented altitude coverage of electron namical processes in which waves, tides, winds, and tur- density profiles in this important region. Another im- bulence generated in the lower atmosphere interact with portant part of EISCAT_3D’s mission will be to provide those generated above (Vincent 1984). Adiabatic cooling a continuous monitor of the high-latitude D region, de- of the upwelling air mass produces a temperature mini- scribing its response to forcing by solar radiation, par- mum of −140 °C at the polar summer mesopause (about ticle precipitation, solar proton events, gravity waves, 85 km), making it the coldest region of the Earth’senvir- meteoric input, and cosmic rays. The latitude and longi- onment (McIntyre 1989). This interplay of chemistry and tude coverage of EISCAT_3D will also allow the spatial dynamics is particularly manifested in the phenomenon of effects of energetic particle precipitation on the atmos- mesospheric thin layers, which show how both processes phere to be investigated. can affect this region of the upper atmosphere (e.g. Ecklund As well as passively observing the changing chemistry and Balsley 1981). of the mesosphere, the EISCAT facilities have the ability The strong scattering layers often seen in the summer to test our models of the system by perturbing the mid- mesosphere, known as polar mesospheric summer dle atmosphere in a known way through active heating echoes (PMSE), are believed to occur due to the forma- experiments (Rietveld et al. 1986; Kero et al. 2000). In- tion of coherent plasma structures in the presence of creasing the temperature of the D region plasma reduces charged particles of nanometre size (Cho and Kelley its recombination rate and increases the formation of 1993). They are formed by the interaction of aerosols, negative ions, allowing chemical models to be tested dust, and ice particles, and are believed to slow down (Ennel et al. 2005). The availability of an EISCAT heat- the diffusion of electrons. PMSE layers often occur in ing facility close to EISCAT_3D, capable of modifying conjunction with the related phenomenon of noctilucent the ionospheric D region, would be a very important clouds (Wälchli et al. 1993; Cho and Röttger 1997; Rapp complement to the new radar, providing an unparalleled et al. 2003b). Sometimes they appear as homogeneous capability for conducting controlled experiments in the single layers, while at other times they can be highly middle atmosphere. structured into multiple braids or sheets (see Fig. 7). Beam focusing of the radar itself could be used to This vertical structuring is believed to be due to atmos- cause ionisation and even breakdown in the strato- pheric temperature fluctuations caused by atmospheric sphere, mesosphere, and lower ionosphere. Such waves (Röttger 2001). Data from existing radars have McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 10 of 63

Fig. 7 PMSE layers observed by the EISCAT VHF radar (courtesy of C. La. Hoz and J. Röttger) indicated that PMSE layers can also have a significant change, which did not exist in the pre-industrial era horizontal structure in reflectivity and velocity associated (Thomas 1996). In this interpretation, the formation of with waves and turbulence (e.g. Yu et al. 2001). 3D these phenomena is believed to have been promoted by imaging in the mesosphere by EISCAT_3D would play a an increase in the amount of water vapour at the meso- major role in investigating the processes of energy coup- pause, caused by the breakdown of methane and other ling in the atmosphere highlighted by PMSE observations. man-made greenhouse gases (Olivero and Thomas One very interesting idea is that PMSE layers, and the 2001). This idea is still controversial, however, since closely associated phenomenon of noctilucent clouds some results suggest that the incidence of mesospheric (Fig. 8), might be symptoms of man-made atmospheric layers has not increased in recent years (e.g. Bremer et al. 2003; Smirnova et al. 2010, 2011). Long-period, high-quality observations are needed in order to reveal whether there are any long-term trends in the incidence or strength of these layers, and to uncover their causes. Although PMSE layers often appear almost laminar, the shape of PMSE spectra suggests that the layers may contain smaller scale coherent structures which have not yet been imaged but may be due to sudden changes in refractive index (Cho and Röttger 1997). These small- scale structures may be part of the explanation for the fact that PMSE layers appear to be strongly aspect sensi- tive, especially at the lowest altitudes (Smirnova et al. 2012). The reason for this aspect sensitivity is not well known; it might be due to the shape of the scattering structures themselves or result from polarisation electric fields, which arise from the charging of heavy ions of dif- ferent sizes (Blix 1999). Alternatively, the aspect sensitiv- ity could be due to the presence of waves or to Fig. 8 Noctilucent clouds over Moscow (courtesy of P. Dalin) horizontal structure in the turbulence, which could be McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 11 of 63

investigated by volumetric imaging (Hill et al. 1999). ionospheric heating, as shown in Fig. 9 (e.g. Havnes et al. The ability of EISCAT_3D to rapidly image the meso- 2004; Kavanagh et al. 2006). When heating is applied, spheric wave field will make it one of the key tools for the intensity of these mesospheric layers is immediately studying the anisotropy and horizontal structure of reduced. This weakening of the layers has been ex- PMSE layers. The multi-static nature of the new radar plained as arising from increases in the electron diffusiv- will allow the layers to be imaged simultaneously from a ity or increased Debye length in the heated plasma variety of aspect angles. In principle, this would make it (e.g. Belova et al. 2003). For PMSE layers, it is also possible to observe coherent scatter from PMSE layers often noted that an “overshoot”,orincreaseinecho with one EISCAT_3D site, while another site probes the strength, occurs at the end of the heating period, pos- same volume via the incoherent scatter spectrum of the sibly because the charge state of the aerosol particles background plasma, at an angle where coherent scatter is temperature dependent (e.g. Havnes et al. 2004; is weak or absent. Biebricher et al 2006). The overshoot effect of PMWE The shape of the backscatter spectrum at the VHF fre- layers seems to be smaller, however, suggesting that if quencies selected for EISCAT_3D is sensitive to key dust particles are involved, they must be smaller in small-scale parameters of PMSE layers, such as the grain size than those in PMSE layers. size of the charged aerosol particles, which is of order 5– PMSE and PMWE layers are not the only types of 70 nm (Hervig et al. 2009). The wide 3D coverage afforded anomalous radar echoes which can be measured in the by the new radar will make it possible to measure how the middle atmosphere. Rocket measurements (e.g. Thrane grain size varies laterally, as well as according to height. 1986; Blix et al. 1990; Strelnikov et al. 2009) have shown Measurements so far suggest that the grain size is likely to that other types of fine structure can also exist in the D be gravitationally ordered, with larger particles predomin- region. These are believed to be associated with plasma ating at lower altitudes (Rapp et al. 2003a). It is also known instabilities, but the process responsible for their forma- that PMSE and noctilucent cloud layers are associated tion is disputed, and it is possible that more than one with turbulence, and one of the proposed applications of type of instability may be contributing to these struc- EISCAT_3D is to use the new radar in conjunction with tures. The high-quality continuous data arising from other radars at different frequencies to explore the EISCAT_3D will make it possible to study these rare spectrum of this turbulence and compare it with theoret- varieties of mesospheric plasma structure and identify ical predictions which link the Schmidt number of the tur- their causes. bulence to the grain size of the layer particles (Rapp and Lübken 2003; Rapp et al. 2008). Atmospheric turbulence in the stratosphere Although strong scattering layers tend to be absent out- and troposphere side the summer months, another thin layer phenomenon The stratosphere (between about 12 and 50 km in known as polar mesospheric winter echoes (or PMWE) altitude) contains the ozone layer, which is vital for can be seen at other times of year (Czechowsky et al. 1989; life on Earth, since it absorbs the energetic part of Lübken et al. 2006, 2007; Kirkwood et al. 2002, 2006a, b, the solar UV radiation, preventing it from penetrating Zeller et al. 2006). PMWE layers generally occur at lower totheground.Thestratosphereisaregionofin- altitudes than PMSE (50–80 km) and are less intense, only creasing temperature with height, since absorption being seen by radars when the background electron density of UV by ozone progressively limits the amount of of the mesosphere is high, for example when the solar radiation available to heat the lower altitudes. This X-ray flux is enhanced or energetic particle precipitation is temperature gradient in the stratosphere means that present (Kirkwood et al. 2002). PMWE occur at higher there is relatively little convective turbulence and that temperatures than PMSE. They are also more aspect sensi- horizontal mixing is much more effective than vertical tive and sometimes seem to have a high horizontal propa- transport (Chen et al. 1994; Shepherd et al. 2000). gation speed. They are clearly not the same as PMSE, and Turbulence can still occur, however, due to variations their nature is still somewhat uncertain—there are some in the and wave activity from thunder- suggestions that they may involve charged aerosols (e.g. storms or local relief (e.g. Kirkwood et al. 2010b), Stebel et al. 2004; Belova et al. 2008; La Hoz and Havnes propagating upwards from the troposphere. The 2008), while others have suggested that PMWE layers stratosphere is notable for the strong interaction it might be purely signatures of atmospheric turbulence exhibits between radiative, chemical, and dynamical (Lübken et al. 2006, Brattli et al. 2006) or that they arise effects. For example, variations in the solar UV out- from highly damped viscosity waves generated by the re- put (one of the most variable parts of the solar flection of infrasound (Kirkwood et al. 2006b). spectrum) can modulate the stratospheric temperature A very interesting finding, made by EISCAT, is that via ozone absorption, changing the heat balance and both PMSE and PMWE layers can be modulated by thereby modifying the spectrum of winds, waves, and McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 12 of 63

Fig. 9 EISCAT data showing the suppression of PMSE layers (left) and PMWE layers (right) by ionospheric heating, which may give clues to the composition of these layers. The left-hand panel shows data first reported by Chilson et al. (2000). The right-hand panel is taken from Kavanagh et al. (2006) tides which is able to propagate between the upper activity. In the troposphere, the air temperature falls and lower atmosphere (e.g. Hartmann 1981). steadily with decreasing height until the tropopause is The stratosphere can exchange energy with both reached, and the temperature begins to increase with the troposphere below and the mesosphere above. A height mainly due to the increased trapping of solar ra- striking example of this type of interaction is shown diation, by stratospheric ozone. The troposphere ex- in sudden stratospheric warming events, which occur changes energy with the stratosphere by dynamical when the circulation of the polar vortex is disrupted, processes such as wind and wave activity and by the re- slowing down or even reversing the direction of the radiation of solar energy back to the layers above. A vortex winds during a period of a few days (Labitzke well-known phenomenon is the trapping of solar radi- 1972, 1981; Simmons 1974). These events occur when ation by greenhouse gases in the lower atmosphere, Rossbywaves,propagatingupwardfromthetropo- which can feed back to the upper atmosphere by re- sphere, grow to large amplitudes as a result of un- ducing re-radiation, producing cooling effects and usual weather patterns. The growing waves interact with changing the dynamics (Lindzen 2007; Archer and the circulation of the stratosphere and, if their upward Pierrehumbert 2011). Chemical effects can produce propagation becomes blocked, they deposit their energy as similar changes in dynamics, as described above, but heat at stratospheric altitudes, causing major changes in which of these mechanisms is most important, the stratospheric temperature and circulation (McIntyre 1982; circumstances under which the different processes Limpasuvan et al. 2004). Recent work, using incoherent dominate, and whether any of them is really signifi- scatter radars as well as other types of radar, has demon- cant in terms of its effects on the lower atmosphere strated that these sudden stratospheric warmings affect are still matters of intense debate. not only the stratosphere but also the dynamics of the The advanced capabilities of EISCAT_3D for digital mesosphere and thermosphere (Kurihara et al. 2010), and beam forming (including post-beam steering within a wide that the effects are not only restricted to the polar regions transmitter beam), radar imaging, and coherent integra- but can be global in scale (Goncharenko and Zhang 2008). tion will make it a superb instrument for stratosphere and The troposphere, at heights up to about 12 km alti- troposphere studies, by exploiting coherent scatter from tude, is the region of the atmosphere in which weather atmospheric turbulence and temperature gradients. VHF phenomena take place. Because this region contains radars are very suitable for this type of work, since the some 75 % of the Earth’s atmosphere, it is by far the radar cross-section of naturally occurring structures is most important region in terms of its effects on human high at these frequencies, the spatial and temporal McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 13 of 63

coherence of the backscatter is large, and the spectrum is inputs to understanding the coupling processes which narrow, making it easily detectable (Rüster et al. 1998). occur there, by observing winds and waves in the meso- VHF radars have measured strong-layered structures sphere (above) and in the lower stratosphere and tropo- in the troposphere and lower stratosphere (e.g. Figs. 10 sphere (below), as well as the energetic particles and 11), which can be complicated and dynamic reaching the mesosphere, and looking for evidence of (Kirkwood et al. 2010b). Although the structures to fine-scale structure which would indicate the presence some extent map the temperature structure of the of turbulent mixing. Measuring the dynamics of the at- lower atmosphere, several of its properties, such as its mosphere above and below the stratosphere provides in- large-scale horizontal variations and the factors gov- formation on which waves are being dissipated at erning its time development, are not well known. stratospheric heights, while observing energetic particles Computer models of developing atmospheric turbu- constrains the modelling of chemical processes which lence have been made, based on processes such as can modulate ozone. Combining EISCAT_3D data with the Kevin-Helmholtz instability and incorporating the other inputs such as satellite observations of strato- effects of thermal and viscous dissipation (e.g. Fritts spheric dynamics and photometric measurements of et al. (2009a,b), but these need to be verified with ozone content will lead to an improved understanding of better observations. Also, the mechanisms of non- the energy balance in the polar middle atmosphere. turbulent scatter (Fresnel scatter) are poorly under- While it is clear that the troposphere can strongly in- stood (Kirkwood et al. 2010a). fluence the stratosphere, via the strong flux of upward- Above the lower stratosphere, however, the strength propagating waves generated at low altitudes, the idea of radar echoes decay rapidly. The viscous sub-range that the stratosphere might be able to impose some form limit for EISCAT_3D wavelengths is reached between of downward control on the troposphere has been much 30 and 40 km (e.g. Kato 2005), although, in some less accepted, until recently (Kuroda and Kodera 1999; conditions (e.g. solar proton events), echoes should be Baldwin and Dunkerton 1999; Song and Robinson 2004). detectable even in this gap. Above this altitude, there If stratospheric effects on the troposphere were shown may be little for EISCAT_3D to scatter from, until to exist, this would provide an extremely interesting link the ionisation of the D region begins to be noticeable in a chain of processes connecting the upper atmosphere between 70 and 80 km. The altitude region between to effects on weather and climate, however indirect that 30 and 70 km is known as the “radar gap”, and ob- connection might be. servations here may still have to rely on other tech- Since the beginning of the present century, it has be- niques, such as lidars or satellite limb sounders. come clearer that there are mechanisms which can allow Although EISCAT_3D will probably not be able to ob- downward control of the troposphere. A pair of well- serve the radar gap, the new radar will make important known papers by Baldwin and Dunkerton (1999, 2001)

Fig. 10 ESRAD radar observations of waves in the lower stratosphere and turbulent structures in the lower troposphere (courtesy of S. Kirkwood) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 14 of 63

Fig. 11 Data from the SOUSY Svalbard Radar, showing stratosphere-troposphere exchange (STE) of energy and momentum, arising from tropopause folding (courtesy of J. Röttger) showed strong evidence for stratospheric precursors of the stratosphere and troposphere which is, as yet, not anomalous tropospheric weather patterns, leading to a well understood. vigorous debate on how such a sequence could arise. A clearly established link between upper atmosphere Three possible mechanisms have been suggested. Firstly, chemistry and dynamics with variability in the tropo- the mean flow of the lower stratosphere can impose a sphere might have significant implications, including “refractive” effect on the flux of upward-propagating providing an explanation for apparent solar cycle signa- Rossby waves (e.g. Hartmann et al. 2000; Limpasuvan tures seen in tropospheric circulation and suggesting and Hartmann 2000) while Rossby waves can even be possible links between long-term variations in strato- completely reflected from higher stratospheric altitudes spheric ozone or greenhouse gases and changes in the (e.g. Perlwitz and Harnik 2003, 2004). In this way, lower atmosphere. However, the extent of downward changes in the stratosphere may be able to modulate control is still a matter of vigorous debate, and great how much of the lower atmospheric wave flux the care will be needed in interpreting the data. stratosphere can accept. Secondly, changes in the poten- There has been considerable recent debate about the tial vorticity of the stratosphere, resulting from wave- idea that geospace processes can influence the tropo- induced forcing, may be able to change the winds and sphere directly, without being mediated by the meso- temperatures of the troposphere, allowing information sphere and stratosphere. One controversial proposal is to propagate downwards across the tropopause (e.g. that cosmic rays can promote cloud formation, by Haynes et al. 1991; Holton et al. 1995). Thirdly, there is modulating the electrical conductivity of the lower at- evidence that when stratospheric wave amplitudes are mosphere. In this scenario, the ionisation of air by cos- large, the interaction between waves and mean flow in mic rays has been proposed to impart an electric charge the stratosphere can generate perturbations sufficiently to aerosols, encouraging them to clump together in large that they can propagate downwards to tropo- groups large enough to form cloud condensation nuclei spheric heights. In addition to these dynamical coupling (CCN). Since the cosmic ray flux varies in antiphase mechanisms, there may be chemical coupling between with the solar cycle, and is probably also modulated by McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 15 of 63

longer term trends in solar activity, the possibility of CCN Short- and long-term changes in the upper atmosphere creation by cosmic rays has been proposed as a direct link The middle and upper atmosphere exhibit a huge between solar variability and tropospheric climate amount of natural variability, on timescales from sub- (e.g. Svensmark and Friis-Christensen 1997; Svensmark seconds to decades. Reasons for this variability include 1998; Marsh and Svensmark 2000a, b). The most recent changes on the Sun, solar wind forcing, plasma instabil- evidence suggests, however, that the observed cosmic ray ities, turbulence, wave-wave interaction, and wave- variability is not capable of modulating the number of particle coupling. Because the atmosphere is highly non- CCNs anything like strongly enough to produce a major linear and coupled, the causes of such variability are not effect on global tropospheric cloud cover (Pierce and easily understood, and even the best physics-based Adams 2009; Kulmala et al. 2009). More recent versions models cannot reliably predict the evolution of the high- of the theory have proposed modifications, such as lim- latitude upper atmosphere, for even the largest scales, iting the connection to a linkage between cosmic rays on timescales of more than a few days (Scherliess et al. and stratospheric (as opposed to tropospheric) clouds 2009; Schunk et al. 2011). (Erlykin et al. 2010). It is believed that stratospheric Superimposed on the small-scale variability of the clouds might be more effective than tropospheric clouds, middle and upper atmosphere are some long-term in terms of their effects on climate, since stratospheric trends, whose causes and extents are subjects of vigor- clouds re-radiate at a much lower temperature than ous debate. Roble and Dickinson (1989) and Rishbeth clouds in the lower atmosphere. and Roble (1992) were among the first authors to draw Although EISCAT_3D is primarily intended as an attention to the fact that, as the lower atmosphere upper atmosphere radar, its considerable sensitivity will warms due to increased heat trapping by greenhouse also make it useful for observations up to an altitude of gases, the reduction in heat re-radiated to the upper around 30 km, crossing the boundary between the atmosphere should cause the mesosphere and thermo- troposphere and the stratosphere. EISCAT_3D has a dif- sphere to cool and contract.Usingthebestthermo- ferent frequency than most other lower atmosphere ra- sphere/ionosphere models available at the time, they dars, and its troposphere/stratosphere observations predicted that a doubling of greenhouse gas concen- would therefore be complementary to data obtained by trations in the lower atmosphere would reduce the air other radars in northern Scandinavia, as well as by satel- density at 300 km by up to 40 %, lowering the peak lites and balloon-borne sounders. The volumetric cap- heightoftheionosphereby15–20 km and reducing abilities of EISCAT_3D can also be used at these the thermospheric temperature by 40 K. This would altitudes, for example to make measurements of wind have some significant effects, for example in increas- vorticity on scales of about 10 km. This could be an im- ing the operational lifetime of satellites because of a portant parameter for meteorological studies, since it reductioniniondrag. has been suggested that vorticity on this scale may trig- Long-term cooling of the mesosphere might lead to in- ger storms (Schumann and Roebber 2010). The physics crease of the incidence of mesospheric thin layers (e.g. of the low-altitude scattering mechanism at these fre- Gadsden 1990, 1998). Analysis carried out by Bremer quencies is also an interesting issue which should be et al. (2006) suggests that although there may be some studied, including issues such as the radar cross-section signs of a trend towards brighter noctilucent clouds and a and the spectral index of atmospheric turbulence and its longer PMSE season, the statistical evidence is presently scattering properties, such as aspect sensitivity. too weak to confirm this as being a significant trend. Also, If a high-frequency (HF) heater is available, as is Kirkwood et al. (2008) and Pertsev et al. (2014) find no planned to be the case for the EISCAT_3D system, it statistically confident long-term trend in moderate or should be possible to create artificial periodic irregu- bright NLC occurrence. larities (APIs) by forming a standing wave between The recent protracted solar activity minimum also the ground and the lower ionosphere. Variations of seems to have resulted in an unprecedented contraction refractive index occur at the nodes of the standing of the Earth’s upper atmosphere (Qian et al. 2008; Tulasi wave, and these can be used as targets for atmos- Ram et al. 2010, Solomon et al. 2010). Combining volu- pheric scattering. This technique has been demon- metric observations of ionospheric velocity made using strated with the existing EISCAT systems (Rietveld EISCAT_3D with simultaneous observations of thermo- and Goncharov 1998) but has not been fully exploited spheric neutral wind, using e.g. a scanning Doppler with the existing EISCAT radars because of their lim- imager, will provide a unique possibility to measure the ited suitability for lower atmosphere work. However, density of the thermosphere in the E and F regions (e.g. the API technique provides the ability to partially Kosch et al. 2011a and references therein). Changes in close the radar gap because observations are possible neutral density, affecting the of space debris ob- from as low as 55 km (Belikovich et al. 2002). jects, can be very dangerous because it makes it more McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 16 of 63

difficult to predict their position, making it very import- It should be noted that, regardless of the sensitivity of ant to secure better predictions by improving the EISCAT_3D, a combination of measurements and mod- current models (Klinkrad 2010; Lewis et al. 2011). EIS- elling will always be needed to recover the parameters of CAT_3D will be very valuable in this respect, since it the thermosphere. In general, the parameters needing to can not only measure the underlying atmospheric pro- be modelled are the neutral density, composition and cesses but also observe the debris objects themselves collision frequencies for energy, and momentum transfer (see “Space debris”). between ionised and neutral species. Reliable estimation Multi-decadal data sets from several regions of the of these can be complicated at high latitudes, especially world, including northern Scandinavia, do indeed appear during active geomagnetic conditions, when the com- to show a persistent decrease in the height of the iono- position and dynamics of the upper atmosphere are per- spheric layers, possibly due to upper atmosphere cooling turbed by processes such as Joule heating and auroral (Keating et al. 2000; Akmaev et al. 2006; Emmert et al. precipitation. Because of this, a close relation between 2008, 2010). An example is shown in Fig. 12. Trends of EISCAT_3D and modelling will be needed, allowing ion temperature at the middle latitudes have been also model and radar results to interact to fully utilise the investigated with Millstone Hill and Saint Santin IS synergy between them. “Ionospheric modelling” contains radar data (Holt and Zhang 2008; Donaldson et al. 2010; some further thoughts in this direction. Zhang and Holt 2013). However, there have been several discrepancies between these observations and estimates Space and plasma physics based on numerical simulations (e.g. Qian et al. 2011; Background Lastovicka et al. 2012). In a major study using EISCAT, The stream of charged particles, the solar wind, blows Ogawa et al. (2014) have shown a cooling trend of 10– continuously from the Sun. When it hits the Earth’s 15 K/decade near the F region peak (220–380 km alti- magnetosphere, it is compressed to a radius of about tude), whereas above 400 km, the trend appears absent 70,000 km on the dayside by the pressure of the solar or may even indicate warming. Further evidence, from wind and stretched into a long tail extending at least studies carried out around the world, indicates the effect 3.5 million km into space on the nightside of the Earth. may not behave in the same way at all locations, leading Some of the energy carried by the solar wind penetrates to speculation that another driving mechanism might be inside the magnetosphere, and a significant part of it dis- involved—perhaps the long-term weakening of the sipates eventually in the Earth’s atmosphere. The activity Earth’s magnetic field or long-term trends in cloudiness of the Sun varies on short (minutes) to long (millions of over different regions of the Earth. This suggests a more years) timescales. Perhaps the most well known is the complicated system than first thought, with its conse- 11-year sunspot cycle; as the number of sunspots in- quent impact on satellite operational lifetimes. creases, so does solar activity. During high solar activity,

Fig. 12 Almost 50 years of data from the ionosonde in Sodankylä, , showing an apparent long-term decrease in the height of the F region (courtesy of Dr. T. Ulich, extending the series of observations first reported by Ulich and Turunen (1997)) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 17 of 63

events called coronal mass ejections (CMEs) often take 1986). A great advantage of the Millstone Hill radar is place, which suddenly and violently release bubbles of its ability to cover a large area by azimuthally scanning gas (plasma) and magnetic fields from the solar atmos- with the dish antenna at very low elevation. The large phere. Large and fast CMEs can approach masses of array of SuperDARN HF radars also look at low eleva- 1013 kg and velocities of 2000 km/s. Earth-impacting tion, probing a medium size area with a spatial reso- CMEs can result in significant geomagnetic storms. Dur- lution of about 45 km in range and time resolution of ing the declining phase of the solar cycle, the high-speed 1–2 min. The radars have proven to be very useful for solar wind emanating from the coronal holes runs into studying the medium-scale dynamic convection patterns, the slower solar wind, and the interaction leads to a in specific on the dayside. However, on the nightside, compression of the plasma and magnetic fields, forming SuperDARN often cannot measure any drifts, probably corotating interaction regions (CIRs). Those have been because its signal gets absorbed during conditions of found to cause also geomagnetic storms and sub-storms. . The difficulties with the HF radar technique for CMEs and CIRs affect the ionosphere mainly by the nightside auroral studies may also be related to the com- process of magnetosphere-ionosphere interaction, whereas plex dynamics of the nightside aurora and its structuring some other effects from the Sun like solar energetic parti- down to much smaller scales. The incoherent scatter cles (SEPs) and radiation bursts at different wavelengths technique at VHF or UHF frequencies combined with can affect directly the ionosphere. volumetric imaging will not only allow the resolution of Observations by the EISCAT_3D radar can hence medium-scale plasma convection under all conditions monitor the direct effects from the Sun on the but also measure electron densities and electron and ion ionosphere-atmosphere system as well as those caused temperatures. by solar wind-magnetosphere-ionosphere interaction. In Based on statistical analysis of HF radar measurements addition, EISCAT_3D can be used to remotely sense the (e.g. SuperDARN), a convection model as a function of magnetosphere. Because the magnetosphere occupies IMF parameters has been produced (Ruohoniemi and such a huge volume of space around the Earth, it is very Greenwald 2005). Several other large-scale models of difficult to observe its large-scale behaviour with in situ global convection exist (see two examples in Fig. 13). spacecraft. However, because of the Earth’s converging The standard data products in these models are typically magnetic field geometry, the high-latitude ionosphere averaged over 1° of latitude (about 110 km). However, acts as an excellent screen for the magnetosphere; pro- scale sizes of a few tens of kilometres and below are im- cesses occurring over hundreds of thousands of kilo- portant, since those represent naturally occurring scales metres in magnetospheric space are projected down into in the ionosphere that can be visually seen as auroral a few hundred kilometres of the ionosphere—a volume arcs and their sub-structure. Indeed, narrow channels of which is feasible for mapping by an imaging radar. enhanced flows are often seen in the ionosphere in the As discussed above, the high-latitude ionosphere is vicinity of auroral forms (e.g. Oksavik et al. 2004). strongly influenced by processes in near-Earth space. In In the magnetotail, satellite measurements have shown addition, the ionospheric plasma can be pumped by that Earthward plasma flow in the is domi- powerful radio waves at frequencies close to the iono- nated by transient fast flows in ambient plasma convec- spheric plasma frequency. Experiments utilising the tion (e.g. Baumjohann et al. 1990). The fast flows have heating facility can be used to study plasma physics in a been observed to appear as bursty bulk flow (BBF) natural laboratory, the ionosphere. events on a timescale of 10 min, with peak velocities about one order of magnitude above the average convec- Plasma convection and multi-scale coupling tion velocities (Angelopoulos et al. 1994). In the high-latitude ionosphere, plasma flows and 3D The BBFs have been observed to be associated with currents are created by the interaction between the solar north-south-extending auroral structures, also known as wind and the terrestrial magnetosphere. The large-scale auroral streamers in the ionosphere (e.g. Henderson plasma convection in the ionosphere is relatively well et al. 1998; Kauristie et al. 2003). However, the details of understood and modelled. At high latitudes, the strength the flow pattern are beyond the resolutions of present and size of the convection pattern, quantified by the so- observing facilities. It has been suggested that the major called cross polar cap potential, are mainly controlled by part of Earthward transport of plasma and magnetic flux the interplanetary magnetic field (IMF) carried by the in the magnetotail actually takes place in the form of solar wind. Both EISCAT and the Millstone Hill incoher- BBFs. There is clear evidence that BBFs are initially ent scatter radar have made important contributions to formed by magnetic reconnection (Øieroset et al. 2000), our understanding of this global convection at both high but this does not explain how they can penetrate so deep and middle latitudes, and one of the available models is into the inner magnetosphere. A theoretical model asso- based on the data from the latter facility (Foster et al. ciating BBFs with underpopulated flux tubes or “plasma McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 18 of 63

Fig. 13 Global convection models: Weimer model (left) and JHU-APL convection model based on SuperDARN measurements, shown as vectors, and IMF conditions (right) bubbles” was developed by Chen and Wolf (1993), and it As in situ measurements of localised transient events is supported by many in situ measurements, e.g. by the in the vast magnetotail can only probe a spatially and Cluster satellites (Nakamura et al. 2005; Walsh et al. temporarily limited part of the phenomenon, therefore 2009). An essential part of the bubble model is the po- ionospheric observations play a key role (Fig. 14). Hence, larisation electric field inside the bubble and associated we need coordinated studies by magnetospheric satellites shear flows and field-aligned currents. and high-resolution volumetric measurements of vector

Fig. 14 Auroral streamers in low-elevation northward-looking EISCAT VHF data. Streamers start from the vicinity of the polar cap boundary

(solid continuous near-horizontal line) and propagate equatorward. Panels from top to bottom: beam-aligned ion velocity, Te, Ne,andTi (Pitkänen et al. 2011). Volumetric multi-static observations are needed to resolve the widths and orientations of the Ne structures as well as associated plasma flows McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 19 of 63

plasma velocities provided by EISCAT_3D. Ground- storms often follow from the interaction of a fast solar based optical observations from the EISCAT_3D volume wind stream or an interplanetary . Mag- would be an important supplement. netic storms typically last from about 12 h to a few days. Storms are characterised by the formation of an intense Sub-storms, storms, and satellite coordination encircling the Earth with current peak at The steady plasma convection in the high-latitude iono- about 4 RE, i.e. well inside the geostationary . The spheres is frequently disturbed, especially on the night- ring current is populated both by efficient convection side of the globe. The reason is a phenomenon known and injection of plasma sheet particles into the inner re- as a magnetospheric sub-storm (Akasofu 1964). Mag- gion and by strongly enhanced ion outflow from the netospheric sub-storms begin with a growth phase, when ionosphere. While sub-storms can occur without mag- a part of the energy derived from the solar wind is netic storms, almost all storms also include sub-storm stored in the magnetotail. The energy is a consequence activity (e.g. Pulkkinen 2007). Space weather phenomena of effective coupling between the IMF carried by the (see “Space weather and service applications”) often ac- solar wind and the Earth’s magnetic field; this process is company (strong) magnetic storms. most efficient when the IMF has a southward compo- Since such a large portion of the near-Earth space and nent. The expansion phase of the magnetospheric sub- upper atmosphere is involved in the disturbance pro- storm explosively releases the stored energy causing duced by a magnetospheric sub-storm or magnetic large-scale changes in the magnetosphere as well as in storm, simultaneous multi-scale observations would be the ionosphere: particle acceleration and precipitation, needed both in space and in the ionosphere. ESA’s Clus- fast plasma flows, enhanced field-aligned currents, en- ter mission was the first multi-satellite mission to ad- hanced ionospheric electrojets, and spectacular auroral dress the question of resolving temporal and spatial displays. During the expansion phase, intense and time- ambiguity in the near-Earth space by using four satel- varying electric fields and currents exist in the iono- lites. Cluster had also an extensive ground-based coord- sphere and those can cause power grid blackouts and ination programme (e.g. Opgenoorth and Lockwood damage to transformers. The whole duration of a sub- 1997; Amm et al. 2005). Another ongoing multi-satellite storm, including growth, expansion, and recovery phases, mission, the NASA Themis mission, is specifically dedi- is typically a couple of hours. cated to study sub-storms. Themis is supported by an ex- The question of which processes control these very dy- tensive number of ground-based observatories in Canada, namic releases of stored energy in the magnetotail con- each including a magnetometer and an all-sky camera. In tinues to be controversial. It has been unclear whether a the future, new multi-satellite missions are expected. The sub-storm starts with the formation of the near-Earth Magnetospheric Multiscale Mission (MMS) is a NASA mis- neutral line (NENL) in the magnetotail at 20–30 Re or sion to study the Earth’s magnetosphere using four identical by a disruption of cross-tail current in the near-Earth spacecraft flying in a tetrahedral formation, building upon magnetotail at 10 Re (e.g. Angelopoulos et al. 2008) and the successes of the ESA Cluster mission. The MMS mis- whether the triggering mechanism is internal to the sion was launched successfully on March 12th 2015. magnetosphere or externally controlled, e.g. by varia- Phased array incoherent scatter radars provide us with tions in the solar wind properties. It is not known what comprehensive ionospheric data locally and over medium effect the state of the magnetosphere has in producing a and small scales. EISCAT_3D is designed to create this particular response mode (e.g. sub-storm, pseudo- opportunity in the Scandinavian sector complementing breakup, or steady magnetospheric convection). Mass the existing phased array incoherent scatter radars PFISR loading of the plasma sheet by ionospheric oxygen may in Alaska and RISR in Canada, but providing multi-scale have a dramatic effect in the tail, and eventually on the and multi-static observations of plasma parameters, in- dayside, when convection transports plasma to the day- cluding ionospheric conductivities and electric fields, side reconnection region (McPherron et al. 2008). It has which can be used to calculate the 3D currents. The also been suggested that ionospheric conductivities most important asset of IS radars is the measurement of could play an important role in allowing the currents to all the important plasma parameters, not obtainable by close via the ionosphere. In addition, the whole nature any other single measurement techniques. of reconnection of magnetic field lines in space plasmas is a subject of intense theoretical and observational re- Auroral dynamics and NEIALs search (e.g. Eastwood et al. 2010). During magnetospheric sub-storms and magnetic Magnetospheric sub-storms require a period of en- storms, the auroral oval is wide and aurora can be seen hanced energy input (southward interplanetary field) even at mid-latitudes. However, auroras are present con- from 30 min to 1 h. If the energy input continues signifi- tinuously and the auroral oval in the nightside iono- cantly longer (>3 h), a magnetic storm develops. Such sphere is most of the time located within northern McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 20 of 63

Scandinavia. Since this area contains a lot of different kinds of ground-based instrumentation including auroral cameras (e.g. ALIS and MIRACLE networks) and pho- tometers, magnetometers, riometers, tomographic satel- lite receivers, etc., it represents a unique location on the globe. In addition, two rocket launch sites (Esrange and Andøya) are located in the area. Evidence for multiple scales in aurora come from sat- ellite and ground-based measurements. The outer scale is the so-called inverted-V structure, typically about 100 km wide when measured by a low-orbiting satellite. Ground-based optical and radar measurements often see auroral arcs with widths of some tens of kilometres (e.g. Knudsen et al. 2001). Structures at 1 km and 100 m scales are also seen (Partamies et al. 2010), and ground- based optical measurements have revealed arc widths down to tens of metres (Maggs and Davis 1968). Large- and medium-scale arcs are often associated with a potential difference between the ionosphere and the Fig. 15 ASK auroral camera measurement at the edge of an auroral magnetosphere, accelerating the electrons into the iono- arc (5 km x 5 km f-o-v at 100 km) revealing a small-scale vortex. The sphere (Mozer et al. 1980; Carlson et al. 1998). How the white circle is the EISCAT UHF beam at 100 km altitude, while the red arrow shows the electric field direction (courtesy of Dr. H. Dahlgren) potential drop develops and to which magnetospheric pro- cesses it is related continues to be unclear (e.g. Borovsky 1993). The origin of narrow auroral arcs is even less understood, even though it has been suggested that measurements of plasma velocities and other plasma Alfvén waves could account for some of the structures parameters by EISCAT_3D are needed. (e.g. Keiling 2009). Intimately connected to structured aurora are the The past studies utilising optical, radar, and satellite small-scale structures known as NEIALs (naturally en- observations have helped to establish the typical electro- hanced ion acoustic lines). This plasma physical dynamic structure of medium-scale (width of some tens phenomenon is characterised by transitory dramatic in- of kilometres) auroral arcs (e.g. Marklund 1984; Aikio creases (four or five orders of magnitude) in the inten- et al. 1993, 2004). However, small-scale structures are sity of the upshifted, downshifted, or both ion acoustic more challenging to measure. Satellite and rocket flights lines observed by EISCAT. Since NEIALs are more co- over these structures give snapshots with a timescale of herent than the normal incoherent scatter spectrum, less than a minute, and the measurement is 1D. The they can be studied by means of interferometry, i.e. conventional one-beam radar measurement suffers from using the relative phases of received signals at multiple space-time ambiguity: it sees only a part of the auroral antenna elements. A result with the two electronically structure at a time. Since the beam width of the current scanning radar (ESR) antennas has been that the en- EISCAT UHF radar is about 2 km in the E region and hanced echoes originate from very localised regions 6 km at 300 km, the auroral structures often fill only (300 m perpendicular to the magnetic field at 500-km partially the radar beam and hence the spectrum of aur- altitude) with varying range distribution and with high oral plasma cannot be correctly estimated. time variability (200 ms) (Grydeland et al. 2004). It has Small-scale structures observed by advanced ground- been reported that NEIALs occur during periods of red based optical TV cameras include also rapidly moving aurora, corresponding to a high flux of low-energy parti- (several km/s) vortices (see Fig. 15) as well as black aur- cles (Collis et al. 1991) and to rayed aurora (Grydeland ora (e.g. Gustavsson et al. 2008; Dahlgren et al. 2010, et al. 2003, 2004; Blixt et al. 2005; see Fig. 16). The ori- and references therein). Black auroras are structures gin of NEIALs is unclear; the possible explanations in- within diffuse aurora with lower luminosity, which can clude current-driven instabilities, driven by strong field- appear as east-west-aligned arc segments or patches, aligned currents; the parametric decay of Langmuir with a typical size order of one to a couple of kilometres. waves (Forme 1993); and possibly related to BBELF They may also exhibit shear or vortices. These small- (broad-band extremely low frequency) waves (Michell scale structures are the projections of some (unknown) et al. 2008; Ogawa et al. 2011). Due to the short occur- plasma processes in the magnetosphere. To test theories rence time and localised occurrence, high time reso- of small-scale arcs and vortices, high-resolution volumetric lution volumetric measurements by EISCAT_3D are McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 21 of 63

At high latitudes, large-scale (hundreds of kilometres) structures in electron density in the F region plasma often exist. One such structure is the ionospheric trough, which is a large-scale depletion in the F region plasma with a width in the latitudinal direction of the order of 5°–10° (for review, see Rodger et al. 1992). The term mid-latitude trough is used for a density depletion with the poleward edge collocated to the equatorward edge of particle precipitation, but it has been suggested that the mid-latitude and high-latitude troughs are prac- tically the same phenomenon extending from mid- latitudes to the polar cap (Whalen 1989). The location of the trough depends on geomagnetic activity and local time. To study the trough with a present EISCAT sys- tem, the UHF radar beam has been scanned in the me- Fig. 16 Rayed auroral arc by 23ox31o auroral camera and the ESR beam mapped to 105 km (yellow circle) when NEIALs were observed ridional or azimuthal direction (e.g. Voiculescu et al. (Grydeland et al. 2004) 2006, 2010; see Fig. 17). Different mechanisms have been proposed for the generation and evolution of the trough (see e.g. Nilsson et al. 2005; Voiculescu et al. 2010). The structure of increased F region electron density needed to study their occurrence, their relation to re- compared to the surrounding plasma is known as the gions of auroral precipitation, and, finally, their cause. tongue-of-ionisation (TOI), in which ionisation from sub-auroral latitudes on the dayside is drawn antisun- Structures and boundaries in the ionosphere ward by the high-latitude plasma convection pattern to- In addition to aurora (associated with increased electron wards the polar cap (Valladares et al. 1994; Pryse et al. densities inside arcs), many other kinds of plasma struc- 2009). At high F region altitudes, the lifetime of the tures exist in the ionosphere. As mentioned in the previ- plasma is sufficiently long for the ionisation to be drawn ous section, one such large-scale structure is the auroral through the polar region and into the nightside sector, oval itself. Both the poleward and equatorward boundar- where it can sometimes be segmented into the polar cap ies of the oval are important boundaries in space as well patches which affect communications and position find- as in the upper atmosphere. Several statistical models ing at high latitudes (see “Space weather and service ap- exist to predict their locations, but instantaneous posi- plications”). In some cases, it has been observed that the tions may still vary a lot. Experimentally, the oval can be tongue-of-ionisation forms the poleward wall of the imaged by high-altitude polar-orbiting satellites. How- main ionisation trough (Middleton et al. 2008). Foster ever, at the moment, no such satellite is in operation. et al. (2005a) utilised several IS radars and the Super- Several other local methods, e.g. utilising satellite par- DARN HF radar network to draw a conclusion that the ticle or radar measurements, exist; specifically, the low- entire low-latitude, auroral, and polar-latitude regions elevation EISCAT VHF radar measurement of electron were fundamentally coupled during the main phase of temperature has been used for detecting the poleward geomagnetic storms, creating a polar tongue of continu- boundary of the oval, which demarcates the polar cap ously streaming cold, dense plasma, along with oxygen (see e.g. the short summary in Aikio et al. 2006; Woodfield ion outflows through the global convection pattern. et al. 2010). Depending on the geographic location of To study the generation mechanisms, evolution, and EISCAT_3D and magnetic activity conditions, the radar coupling between different latitude regions of these could contribute to the detection of either of the bound- large-scale electron density structures, data from a global aries. Ground-based networks of optical instruments network of instruments are needed. Radio receivers are an important complement for these studies. Support using tomographic methods can provide electron density for such measurements ideally requires low elevations maps, usually in 2D, in the future more often also in 3D. to the north and south to maximise spatial coverage. Convection can be studied by HF and IS radar networks. Note that it was a requirement of the EISCAT_3D an- Volumetric EISCAT_3D radar can provide very valuable tenna and array design that the gain reduction should contribution to these studies by measuring the 3D distri- be as small as possible for elevations up to 60° off zen- bution of all plasma parameters and helping to solve ith. The report by Johansson et al. (2014) shows the gain problems related to in situ plasma production and loss. performance that can be achieved for the designs under All the measurements should be combined with model- consideration. ling efforts. At the moment, several numerical models of McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 22 of 63

Fig. 17 Ionospheric trough observed by means of satellite tomography (top) and EISCAT UHF meridian scan (bottom) (Voiculescu et al. 2006) the upper atmosphere such as the Coupled Thermosphere- generation by utilising the multi-static multi-beam Ionosphere- (CTIP) model (Fuller-Rowell plasma velocity measurements. et al. 1987) and the Utah State University time-dependent ionospheric model (Sojka et al. 1981) are available to Currents and energy input into the ionosphere model these persistent, large-scale features of the high- Large-scale eastward and westward electrojets flow latitude ionosphere. within the auroral oval. Their intensity and extent is At high latitudes, the E region electron densities are controlled by conditions in the near-Earth space, which often enhanced by auroral precipitation. Very large are characterised by magnetic activity indices. In most of changes are caused by narrow layers known as spor- the modelling studies, it is assumed that all the horizon- adic E. Sporadic E (Es) layers are thin horizontal tal currents in the ionosphere flow within an infinitely layers with a vertical extent of 0.5–5 km and densities thin sheet at some fixed altitude (usually 110 km) above at least two-three times higher than the background the ground. The thin sheet approximation simplifies the plasma. Their sources are metallic ions, formed by analysis, but it is not always a sufficiently accurate de- ablating meteors. Much of the progress in under- scription of the ionosphere. In reality, the Pedersen and standing high-latitude sporadic E comes from studies Hall conductivities maximise at different E region alti- made with the EISCAT radars (see the review by tudes and hence also the respective Pedersen and Hall Kirkwood and Nilsson 2000). An important gener- currents have different altitude profiles. To calculate the ation mechanism for Es layers at mid-latitudes is the 3D currents in the ionosphere, we need to have a volu- wind shear mechanism, but at high latitudes, the elec- metric measurement of electron density and ion velocity tric field mechanism, vertical convergence of iono- vector. Real altitude-resolved ionospheric measurements spheric plasma by the magnetospheric electric field provide the possibility to describe the vertical closure of (Nygrén et al. 1984), becomes important. current within the ionosphere and to investigate the in- Even though Es layers have been known a long time ductive coupling between the currents. Indeed, model (initially they were identified from ionosonde record- calculations have shown that the internal induction in ings), their horizontal distribution is not well known. the ionosphere may produce significant rotational elec- EISCAT_3D will play an important role in studying the tric fields (Vanhamäki et al. 2007). The induced electric horizontal distribution and evolution of these layers as field is important at local “hot spots”, reaching values well as the role of winds vs. electric fields in layer 20–50 % of the potential electric field present at the McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 23 of 63

same locations. Also, the induced FAC make a large con- the 3D distribution and composition of the upflowing tribution to the total field-aligned currents flowing be- ions and neutrals. Detailed knowledge of these pro- tween the ionosphere and magnetosphere. The effect of cesses is important for model predictions of what will induction phenomena on magnetosphere-ionosphere happen to the composition and density of our atmos- coupling cannot be studied without volumetric vector phere in the long-term future; for example, it could tell measurements from the ionosphere, to be provided by us whether the entire atmosphere will eventually evap- EISCAT_3D. orate and be lost into space. The 3D structure of ionospheric currents is directly re- There are several transition regions for upflowing ions, lated to 3D Joule heating in the ionosphere. Recent esti- for example, from chemical to diffusion dominance at mates indicate that the polar represent a 500–800 km altitude, from subsonic to supersonic flow major sink for the energy arriving inside the magneto- at 1000–2000 km altitude, and from collisional to colli- sphere by the solar wind-magnetosphere coupling. The sionless at 1500–2500 km altitude. EISCAT_3D is one of role of Joule heating has been estimated even as 50– the most suitable instruments to investigate such transi- 60 % (Tanskanen et al. 2002; Ostgaard et al. 2002). The tions because of its wider height coverage (up to about global particle heating has been estimated as 29 % by 2000 km) along the field line. EISCAT_3D will give in- Ostgaard et al. (2002). Finally, the energy goes to the formation on the accurate thermal ion velocity and up- thermosphere and may affect thermospheric dynamics ward flux, whereas thermal ion detectors on satellites and vertical energy coupling in the atmosphere. suffer from the effect of positive spacecraft charging. To evaluate the energy deposited in the 3D volume of Upward ion flow in the polar ionosphere depends on the ionosphere, we need to have a volumetric measure- particle precipitation (increasing electron temperature) ment of electron density and ion velocity vector. Ion and plasma convection (causing frictional heating), but velocity vector measurement together with model ion- observations show that upflow is also influenced by solar neutral collision frequency altitude profile can be used activity, season, and geomagnetic activity. In the iono- to calculate the 3D neutral winds in the E region sphere, upward-flowing ions are seen predominantly in (Heinselman and Nicolls 2008; Nygrén et al. 2011), the polar region. The ESR radar is located near the cusp which are needed in estimating Joule heating. With region where ion outflow and neutral upwelling fre- EISCAT_3D, it would possible for the first time to quently occur (see Fig. 18). EISCAT_3D will be monitor- quantifytheenergyinputintheformofJouleheating ing the auroral oval, which is also an important source and particle precipitation on the spatial scales of aur- for upflowing ions, as evidenced by the Tromsø UHF oral arcs. Joule heating may vary by >100 % on mesos- radar (e.g. Ogawa et al. 2010). patial (~60 km) scales (Kosch et al. 2011a). If small- scale structures (e.g. intense localised electric fields in Ionospheric modelling the vicinity of auroral arcs) are totally absent from the Many existing models describe successfully only average models, this leads to potentially significant underesti- conditions of the non-auroral, non-perturbed iono- mation of Joule heating rates (e.g. Deng and Ridley sphere, like the empirical global International Reference 2007; Aikio and Selkälä 2009). Ionosphere (IRI) model (Bilitza 2001). A number of physics-based models exist, too. The Global Ionosphere- Ion outflows Thermosphere Model (GITM) is a 3D physics-based An important phenomenon of magnetosphere-ionosphere model of the Earth’s thermosphere and ionosphere sys- coupling is the formation of upwelling ion populations in tem (Ridley et al. 2006). GITM is coupled to a large the topside polar ionosphere. These upflows can represent number of models of the high-latitude ionospheric elec- a significant loss of atmospheric gases into interplanetary trodynamics, for example, the assimilative mapping of space and a significant source of magnetospheric plasma, ionospheric electrodynamics (AMIE) technique, and which may also affect the dynamics of the magnetosphere Weimer, Foster, Heppner, and Maynard or Ridley et al. (e.g. Yau and André 1997; Yau et al. 2007). Key processes electrodynamic potential patterns. Input parameters in- for upward ion flows in the topside ionosphere are sug- clude the F10.7 index and solar wind parameters (e.g. gested to be frictional heating, ambipolar diffusion Weimer 2005). The latitude resolution is 2.5°, and longi- driven by a heated electron gas, and transverse ion tude resolution is 5° and output parameters include neu- acceleration produced by plasma waves. Artificially in- tral, ion, and electron temperatures; neutral winds; duced ion upwelling using the EISCAT heater demon- neutral densities; and plasma velocities. Another model strates that ambipolar diffusion driven by the heated relevant to high-latitude electrodynamics is GUMICS, a electron gas is a viable mechanism (Kosch et al. 2010). global solar wind-magnetosphere-ionosphere coupling It is critical to determine the relative importance of the model. Its solar wind and magnetospheric part is based different mechanisms in operation and to understand on solving the ideal MHD equations, and its ionosphere McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 24 of 63

Fig. 18 Both hydrogen ion (H+) and oxygen ion (O+) upflows in the topside polar ionosphere were recently observed with the ESR radar in the vicinity of the cusp. On closed field lines, the H+ became the larger contributor to the upward flux above about 550 km (see Ogawa et al. 2009) part is based on solving the electrostatic current con- difficult to use the kinetic approach at low altitudes. The tinuity equation (Janhunen 1996). Model input are solar core of TRANSCAR is made of two models: a 13- wind parameters and model output in the magneto- moment fluid code that deals with thermal electron and sphere contain plasma density, pressure, velocity, and ion (six species considered) transport above 90 km and a magnetic field vector in space and time, and in the iono- kinetic part that takes care of the ionisation and energy sphere Pedersen and Hall conductances, ionospheric po- deposition resulting from solar illumination and particle tential, ionospheric electric field, precipitation power, precipitation (two yellow blocks in Fig. 19). Joule heating rate, and field-aligned currents in space Two other important components of the model are and time. The above-mentioned global models can be the magnetospheric transport unit that makes the flux accessed by the aid of the Community Coordinated tubes convect (electric field and potential maps needed) Modeling Center (CCMC) (http://ccmc.gsfc.nasa.gov/). and a neutral ionosphere unit (based on the MSIS-90 Global models have usually poor temporal and spatial model) for the neutral-ion interactions mainly. Although resolution. To study the small-scale (tens of kilometres TRANSCAR is not a self-consistent model, it is able to to sub-kilometres) structures that EISCAT_3D will be account for electrodynamic coupling between the mag- able to measure, more detailed models are needed. We netosphere and the ionosphere. Besides the knowledge review below some such models (TRANSCAR, IMM, of the neutral atmosphere, which can be adjusted from and KIMIE), which have been developed in the iono- the MSIS-90 empirical model by calibrating the model spheric group at IRAP (Toulouse, France) in order to during a quiet period (Blelly et al. 1996), it requires in- better understand the high-latitude ionosphere and its put parameters concerning the precipitating particles coupling with the magnetosphere and to compare with and the convection electric field. The precipitation can EISCAT radar data and numerical simulations. In the fu- be given by serendipitous low-altitude satellite passes, ture, the models may be coupled together. while information about the convection can be inferred from HF radar observations (SuperDARN) or a combin- TRANSCAR TRANSCAR (Blelly et al. 1995, Diloy et al. ation of HF radars, magnetometers, and satellites (AMIE 1996) describes the transport of several ionospheric spe- procedure). In the latter case, AMIE actually gives also cies (electrons and six ions) along a magnetic field line the precipitation and the field-aligned currents. Lately, between 90 and 3000 km. Two approaches can be used TRANSCAR was coupled with the IMM (Ionosphere- in order to describe the ionosphere. Either one considers Magnetosphere Model described thereafter) which can the temporal evolution of the distribution functions of a also provide these necessary input parameters. given species (kinetic approach) or one considers the The outputs of this model are compatible with the pa- temporal evolution of the moments of the distribution rameters measured by incoherent scatter radars, so that function (fluid approach). In the case of the Earth’sat- the results of the modelling can be directly compared to mosphere, the complexity of the collision terms makes it the observations. The model can be used as a virtual McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 25 of 63

Fig. 19 Schematic figure of the TRANSCAR model (courtesy of Pierre-Louis Blelly) instrument, which a user can locate almost anywhere on which IMM computes the ionospheric electric field. The Earth. One can modify individually ionospheric parame- logic of the model is presented more precisely in Fig. 21, ters such as the convection electric field, the precipitation, adapted from Peymirat and Fontaine (1994). the neutral atmosphere, and so forth. TRANSCAR has The interaction of the solar wind with the Earth’s been used to model the time evolution of the plasma pa- magnetosphere creates first an electric field in the rameters within a given convecting across the magnetosphere, which induces the convection of the polar cap. This is obviously not observable but TRANS- magnetospheric plasma. During this motion, the plasma CAR makes it measurable (Fig. 20). There is an exciting is compressed such that potential for such a model to be utilised to help analyse/ process existing EISCAT and future EISCAT_3D data a) The plasma escapes from the flux tubes and through a Bayesian approach. Such an approach could po- precipitates in the ionosphere. This precipitation, tentially enable the extraction of parameters not directly mainly due to electrons, increases the conductivity given by the existing analysis method (composition) or by of the ionosphere and generates the first coupling the incoherent scatter theory at all (thermal electron between the magnetosphere and the ionosphere. velocity). b) Pressure gradients are created in the magnetosphere which induce electric currents flowing orthogonally IMM The IMM (Peymirat and Fontaine 1994) is based to the magnetic field. These pressure gradients are on the fluid formalism and computes the motion of the mainly due to the ions, because their pressure is magnetospheric plasma in the inner magnetosphere, tak- larger than the one of the electrons. The divergence ing into account its coupling with the ionosphere. The of these orthogonal electric currents generates 3D equations describing the motion of the magneto- electric currents flowing along the magnetic field spheric plasma are reduced to two dimensions with an lines, the so-called field-aligned currents. They close integration of these equations along the magnetic field in the ionosphere and induce electric currents flow- lines, such that IMM computes the motion of the flux ing horizontally in the ionosphere. They correspond tubes in the equatorial plane of the magnetosphere. to a second coupling between the magnetosphere Similarly, the 3D equations describing the dynamics of and the ionosphere. the ionosphere are reduced to two dimensions, with an integration in altitude, such that the ionosphere is de- From Ohm’s law, one can then compute the electric scribed as a 2D thin layer surrounding the Earth, from field in the ionosphere, which maps to the magnetosphere McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 26 of 63

Fig. 20 Comparison of EISCAT measurement (upper panel) and TRANSCAR simulation (lower panel) (Blelly et al. 2005) along the magnetic field lines which are assumed to be field-aligned currents, and the equipotentiality of the mag- equipotential. This mapping modifies the initial electric netic field lines. field induced by the solar wind and requires that the As mentioned above, the coupling between IMM and modelling of magnetospheric plasma convection has to TRANSCAR has provided several benefits: the IMM take account of the coupling between the magnetosphere providing TRANSCAR with some necessary inputs like and the ionosphere induced by the precipitation, the the convection electric field and the field-aligned current McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 27 of 63

Fig. 21 Schematic of the IMM model (Peymirat and Fontaine 1994) densities and TRANSCAR providing IMM with the with parallel electric fields much lower than those pre- ionospheric conductances. In turn, the coupling with dicted by the fluid theory. IMM facilitates a better interaction between EISCAT_3D and models such as TRANSCAR. EISCAT_3D model joint studies The three models de- scribed above will eventually be combined, by coupling KIMIE KIMIE (KInetic Model of Ionospheric Electrons) IMM and KIMIE to TRANSCAR. The use of such a is a 1D kinetic model developed by Garcia and Forme complete ionospheric model is threefold: (2009). It describes the transport of electrons in a colli- sional and partly ionised plasma, including electron-  A particular EISCAT_3D data set can be modelled electron, electron-ion, and electron-neutral collisions. and a close comparison can be made at any required Collisions between charged particles and between temporal and spatial resolution. charges and neutral particles are treated differently.  Various physical processes can be tested on a Electron-neutral collisions are binary and are described synthetic or real event, by tuning the input by the Monte Carlo method, while collisions between parameters of the model. charged particles are long-range interactions, modelled  Nowcasting and even forecasting of the physical by Fokker-Planck equations. This code was initially de- parameters in a given ionospheric volume could in veloped to study the kinetic effects due to strong principle be performed, provided that sufficiently current density in the Earth’s ionosphere. Its first results realistic inputs (solar extreme UV (EUV) flux, particle showed that very strong current densities result in non- precipitation, and electric field essentially) can be Maxwellian distribution functions and that such cur- supplied by real-time data or empirical models. rents are carried mainly by suprathermal electrons. These suprathermal electrons significantly modify the Although TRANSCAR solves the transport equations local conductivities and may carry very strong currents in 1D along the flux tubes, it is in fact a pseudo-3D McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 28 of 63

model, since it requires the modelling of several flux stations located in northern Scandinavia. The typical tubes as they evolve in space and time, even to make time resolution is about 5–10 s. By inverting vertical predictions for a single point. Simulating a whole iono- profiles of the blue VER at 427.8 nm (N2+ 1NG) with spheric volume is therefore already feasible. tomography-like techniques, fluxes of precipitating elec- trons can be inferred in 2D from ALIS data, for example  The resolution of the modelling far surpasses what across and along an auroral arc. The current configur- can be observationally achieved by the present ation of EISCAT only allows precipitating electron fluxes EISCAT system. EISCAT_3D’s spatial and temporal to be retrieved in one dimension, along the magnetic resolutions will make it possible to compare data field line. EISCAT_3D will be able to produce 2D maps and simulations at smaller scales, enabling a better of electron fluxes and with a higher time resolution than understanding of the microphysical phenomena. currently provided by ALIS. These electron fluxes can be  Although it is possible in principle to infer ion used as inputs for TRANSCAR/TRANS4 in order to re- composition from incoherent scatter data, this is not produce either electron densities and temperature profiles done routinely for the present EISCAT systems. One obtained from EISCAT_3D measurements, or emission problem is that ion composition effects on the line profiles obtained with ALIS. Such flux measurements incoherent scatter spectrum are to a large extent would also be very useful to test magnetosphere- ambiguous with effects due to ion temperature. ionosphere coupling models and could be compared to Multi-parameter fitting to recover both temperature in situ data from a spacecraft crossing the same mag- and composition is possible but requires very low netic field line at higher altitude. measurement errors in the amplitude comparison function ACF (autocorrelation function). For the FMI electrodynamical models In order to get an accur- present EISCAT radars, these ACF errors are ate picture of ionospheric electrodynamics, several dif- basically determined by signal to noise ratio. The ferent physical quantities need to be observed with good accuracy of EISCAT_3D spatial and temporal resolution. This is most readily rea- measurements, however, will be limited only by lised with large instrument networks, such as MIRACLE, the self-noise of the transmitter codes (Lehtinen which consists of magnetometers, all-sky cameras (ASC), et al. 2014). This, and the possibility to contain and radars situated in northern Europe and operated and estimates from all of the EISCAT_3D sites, owned by the Finnish Meteorological Institute (FMI). Sev- should facilitate much better composition eral other instruments that are not part of the MIRACLE measurements from EISCAT_3D. network also operate in the same area (e.g. the EISCAT  With the tristatic EISCAT UHF system, it is possible incoherent scatter radars and SuperDARN radars). The to determine the three components of the electric FMI group has developed a set of different analysis field, but only at a single point. Having several methods for solving the ionospheric electrodynamical simultaneous beams will be a huge improvement for, parameters based on MIRACLE and supporting obser- among other things, the determination of electric vations (e.g. Amm et al. 2003). field vectors at several points in an ionospheric EISCAT_3D should cover an analysis volume of at least volume. EISCAT_3D will be able to measure electric 300 km × 300 km × 50 km (altitude) and hence will cover field distributions in altitude and horizontally in the typical extent of mesoscale auroral forms for integra- latitude/longitude, which will be a great tion with magnetometer-based studies in connection with improvement for the study and modelling of MIRACLE. The radar observations will need to cover the electrodynamical phenomena, such as the E region, in order to be able to estimate conductances. ionospheric response to ULF waves. Spatial resolutions of at least 20 km × 20 km × 2 km (alti- tude), and temporal resolutions between 10 s and 1 min, Attempts to compare the present EISCAT data with are needed for integration with MIRACLE, but these will predictions from TRANSCAR, IMM, and KIMIE are not be a problem for EISCAT_3D. The parameters that limited by a number of factors. These issues, which can EISCAT_3D could provide as input into modelling, or all be resolved by EISCAT_3D, are the following: which could be compared with model results, include The volumetric imaging capability of EISCAT_3D will electric fields and Hall and Pedersen conductivities. From also be extremely useful for coordinated observations those, horizontal currents could be calculated and vertical with optical networks like ALIS (Auroral Large Imaging currents could be inferred from the divergence of the System, owned by the Swedish Institute of Space Re- horizontal currents. search, ). ALIS is capable of providing 3D volume emission rates (VER) for several emission lines from a SIC The Sodankylä Ion-Neutral Chemistry (SIC) model set of simultaneous optical images obtained from several is a comprehensive 1D (in altitude) coupled model of McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 29 of 63

the chemistry of the middle atmosphere, covering the Many of the topics discussed above, such as NEIALs, mesosphere and lower thermosphere as well as the iono- can also be treated as examples of plasma physical pro- spheric D region. SIC has been developed by the SGO cesses. In addition to naturally occurring phenomena, Aeronomy Group in close collaboration with the Finnish plasma physical phenomena can be excited artificially by Meteorological Institute (Turunen et al. 1996). feeding energy to plasma by radio waves at frequencies The SIC model includes more than 400 reactions of close to natural plasma frequency. EISCAT, through the positive and ions and neutral constituents. Sources of use of the heating facility, has been at the forefront of re- ionisation accounted for are solar EUV, X rays, protons, search into artificially induced instabilities, and EIS- and electrons, as well as galactic cosmic rays. The alti- CAT_3D has the potential to make further very significant tude range is from 20 to 150 km and the typical time discoveries both because its higher gain will allow much resolution is 5 min. Output parameters are ionisation better time resolution of fast coupling processes and be- and dissociation rates, electron density, ion composition, cause of its ability to carry out quasi-simultaneous imaging and odd oxygen, hydrogen, and nitrogen concentrations. of the entire pumped volume, leading to the possibility to SIC can be run either in a steady-state mode, which cal- perform greatly enhanced investigations of spatial and culates the equilibrium concentrations of the modelled temporal effects (see Tables 1 and 2 in the Appendix). components, or in a time-dependent mode where the In the ionosphere, powerful radio waves can excite concentrations are advanced in time using the modified Langmuir turbulence and upper-hybrid resonance, raise Euler method for stiff equations. The steady-state mode the electron temperature greatly (up to 4000 K), and is suitable for estimating the effects of constant forcing trigger ion upflow in the topside ionosphere (Kosch during daytime, whereas the time-dependent mode al- et al. 2010; Rietveld et al. 2003, and references therein). lows us to study effects of diurnal variations or transient Critical for the efficiency of the heating is the angle that events, such as precipitation of energetic particles. the HF pump wave vector makes with the geomagnetic EISCAT measurements are essential in understanding field, but the reasons for this are not clear. Kosch et al. the effect of energetic particle precipitation (EPP) on the (2011b) found that adjusting the pointing direction of neutral atmosphere. Particle precipitation and ion chem- the radar by 1° resulted in a different observation, in istry can affect mesospheric ozone by producing HOx. which the ionospheric radio window in the bottomside EISCAT_3D observations can be used to provide iono- ionosphere was observed at 7°–8° south of zenith, along spheric inputs for the SIC model, yielding with HF radio wave penetration to the topside iono- sphere. The wavelength of the VHF radar is ideally  Continuous observations (long-term characterisation suited to study Langmuir turbulence. Exceptionally of EPP forcing) bright optical emissions attributed to Langmuir turbu-  Latitude-longitude imaging capability (spatial extent lence were observed in self-focusing of the transmitted of EPP) pump wave (Kosch et al. 2004). Artificial optical emis-  D region coverage (HOx/ozone region) sions are normally observed when upper-hybrid reson- ance is stimulated. It has been discovered, using the Plasma physics and active experiments EISCAT heating facility, that the emissions depend The geospace environment forms a superb natural la- strongly on the relationship between the pump fre- boratory for the study of plasma physics, affording a bet- quency and the electron gyro-harmonics (Gustavsson ter vacuum than the best obtainable on Earth as well as et al. 2006; Kosch et al. 2002). The gyro-harmonics, in a measuring volume uncontaminated by vessel geom- turn, depend on the magnetic field strength and the alti- etries or edge effects. Since more than 99 % of the observ- tude in the ionosphere. The question arises of what elec- able universe consists of plasma, passive observations and tron acceleration mechanisms are at play in (what active experiments using the solar-terrestrial plasma appears to be) a new non-linear regime. provide a unique window on processes that are funda- One very interesting aspect of EISCAT_3D’s ability to mental throughout the universe. Since the density and support plasma turbulence experiments arises due to its composition of the rarefied neutral atmosphere varies capability for sub-beam width imaging, which will allow strongly as a function of height, the kind of plasma the small-scale density irregularities produced by the found in the upper atmosphere is also a strong function heating facility to be imaged for the first time. The size of altitude, ranging from charged dust/ice complexes in and distribution of these filaments and the way in which themesosphere(see“Dynamical and chemical coupling their properties evolve during pumping are not well in the mesosphere”), through plasmas dominated by known because existing radars do not have sufficient molecular species and atomic oxygen in the upper iono- resolution to measure them; however, they are important sphere, to hydrogen-dominated plasmas at magneto- to study, for example, because they affect satellite com- spheric altitudes. munications and navigation and because of the McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 30 of 63

possibility that the density cavitations might play a role antenna array, allowing “beat frequency” experiments, in electron acceleration. The temporal focusing of artifi- where the heater frequency is provided by the beat fre- cially induced optical emissions provides indirect evi- quency between the two transmissions. This would be dence for this (Kosch et al. 2007). Last but not least, the an excellent technique to generate some of the very low evolution and self-organisation of the wide spectrum of frequencies presently inaccessible to the Tromsø heating spatial scales of the density filaments teach us about facility. Another interesting property of EISCAT_3D is how plasmas dissipate energy flow in general. To date, the possible use of phase and polarisation flexibility to only one experiment has detected a glimpse of the spatial create “twisted beams” carrying orbital angular momen- organisation of the density filaments, obtained by in situ tum (Leyser et al. 2009), which can interact with angular measurements with a rocket at the mid-latitude Arecibo momentum phenomena in the ionosphere. There are facility (Kelley et al. 1995). This experiment indicated a theoretical predictions of twisted beams of plasma most interesting organisation of the density structure, with waves, specifically of Langmuir waves (plasmons) and a wide range of coupled spatial scales. The observed ion acoustic waves (phonons) (Mendonca et al. 2009a, spatial scale down to below 10 m gives the resolution b), and it has been suggested that such twisted wave transverse to the geomagnetic field that should be aimed groups can be excited in aurora. Twisted beams can also for with EISCAT_3D. However, the physics of the re- be used to detect small-scale plasma flows transverse to sponse of the ionospheric plasma as it is pumped by the the beam that fit within the radar beam cross-section, powerful radio wave is likely to be significantly different at such as vortices and sheared flows. To detect angular high auroral latitudes than at the mid-latitude Arecibo fa- momentum effects of strong plasma flows in the iono- cility, because of the different geomagnetic field geometry; sphere requires high degrees of beam twisting together much more pump energy is dissipated through the forma- with narrow beams (orbital angular momentum mode tion of density filaments at high latitudes. number of 10 or higher with a beam width of 1°). In addition to its use as a radar, EISCAT_3D has the At the bottom of the cascade linking large-scale and possibility to act as a D region heater by focusing its small-scale structures lies the whole area of plasma tur- power into a narrow beam. Although the frequency will bulence, whose understanding and description remain be too high to stimulate ionospheric plasma resonances, one of the most important unsolved problems in clas- the huge effective radiated power (100 s of GW) will sical physics. Turbulent plasma comprises an array of make for an extremely effective ohmic heater of the structures covering a wide range of scale sizes and life- ionospheric plasma, especially in the D region. Such cap- times, all of which are non-linearly created, coupled, ability would be invaluable in the study of dusty plasma, and destroyed. In its turbulent state, plasma reaches a which occurs due to meteoric ablation in the meso- quasi-stationary point, which is, however, very far from sphere. The radar signature of dusty plasma (polar me- thermodynamic equilibrium. The effects of plasma tur- sospheric summer/winter echoes) is modified by heating bulence are extremely important; for example, turbu- the surrounding plasma (e.g. Havnes et al. 2003; Kava- lent diffusion and transportrepresentafundamental nagh et al. 2006), allowing the dust properties such as limitation on the ability to confine energy within grain size and charge to be inferred. Ionospheric studies plasma. of dusty plasma relate directly to astrophysical plasmas, An important but unexplored and difficult type of ex- which are also dusty, but also to climate change because periments are those that attempt interaction of the the mere existence of radar echoes from dusty plasma is transmitted pump wave with natural (free) energy due to increasing mesospheric water content and cool- sources (Leyser and Wong 2009), including pump- ing, both brought on by climate change (Olivero and induced Langmuir turbulence interacting with auroral Thomas 2001). Beam focusing can be used to cause ion- electron precipitation or naturally enhanced ion acoustic isation and even breakdown in the atmosphere (strato- waves. By seeding the ionosphere with Langmuir or ion sphere and mesosphere) and lower ionosphere. Such acoustic waves during appropriate conditions of auroral experiments have the potential to give information on a electron precipitation, energy might be channeled into range of atmospheric properties and constituents that the plasma waves, which might open up for studying are not accessible by natural excitations. This would also new non-linear regimes of plasma turbulence. Further, in allow ULF and VLF electrojet modulation experiments. such experiments, possible feedback of the ionospheric Such low-frequency ULF and VLF transmissions open plasma turbulence on the larger scale system supplying the possibility to study the ionospheric Alfvén resonator the precipitation might be studied. Interferometric mea- (Robinson et al. 2000) as well as cyclotron resonance in- surements are important to study the self-organisation of teractions with radiation belt particles (Inan et al. 2004). the turbulence in the pump-plasma interaction region, in- EISCAT_3D may also allow the flexibility to transmit cluding the possibility of filamentation, which has been different frequencies from different sections of the predicted theoretically for a long time. McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 31 of 63

Solar system research micrometeor orbits with a view to eventually determin- Background ing the 3D distribution of sub-millimetre particles at Even though EISCAT_3D is designed to study the iono- 1 AU with good statistics. However, the UHF system has sphere and atmosphere of the Earth, it can also be used not by any means been ideal for this purpose, since the to study the properties of our solar system. Due to the scattering cross-section is strongly inversely proportional high power and great accuracy, mapping of objects like to radar frequency and hence the number of orbits has the Moon and asteroids can be done. Material in the been relatively small. All the meteors were found to be form of dust and meteors from our own solar system, bound to the solar system (Fig. 22). When plotted along and also from interstellar space, impacts the Earth’sat- the orbital semi-major axis, their orbits showed indica- mosphere continuously. Because of their high power and tions of resonance gaps induced by interaction with the large antenna aperture, incoherent scatter radars can be gas giants Jupiter and Saturn at the positions where extraordinarily good monitors of extraterrestrial dust these gaps have been modelled to appear (Szasz et al. and its interaction with the atmosphere. It is very im- 2008). In addition, measurements from three different portant to make good measurements of the flux of me- look angles have made it possible to study the aspect teoric material entering the upper atmosphere, because angle dependence of the echo strength and thus indir- meteoric dust plays an important role in the chemistry ectly the plasma dynamics around the meteor head and and heat balance of the middle atmosphere and thus fragmentation of individual meteors in the atmosphere forms a very important input into middle atmosphere (Kero et al. 2008a, b). models. In addition, the observations can contribute to EISCAT_3D has good potential to be a powerful instru- work on solar system dust models (e.g. Mann 1995). Last, ment for studies of the effects of the meteor influx as well if wideband receiving capability is retained, the work uti- as the off-ecliptic component of the interstellar dust. At lising interplanetary scintillation can be continued to give the VHF operating frequency of 233 MHz, event rates will information of the structures in the solar wind. be much higher than at 930 MHz, probably by more than an order of magnitude. The multi-beaming capability Meteoroids makes it possible to perform three or more tristatic obser- Meteoroids roam through the solar system with orbits vations at different heights simultaneously, which will in- of all inclinations. Meteoroids (interplanetary or inter- crease the event rate further by a factor of 3 or more. stellar debris) range in size from small asteroids with Another beneficial effect of the lower operating frequency radii of ~10 km down to micrometeoroids with radii is that head echoes will become observable at up to 115- of ~0.1 mm and dust, radii of ~1 m. When a meteor- km altitude. This enables observing interstellar particles, oid enters the atmosphere, it is heated up and it loses which disintegrate at higher altitudes than particles on mass due to ablation. The time duration of visible closed orbits due to their higher velocities. meteors is less than 1 s, and the main ablation takes The Canadian Meteor Orbit Radar (CMOR) survey place from 140 to 70 km. The radar target is pro- (Brown et al. 2008), conducted between 2002 and vided by the coherent reflection from ionised plasma, 2008, measured the orbital parameters of 2.5 million and there are two types of echoes: meteor head and meteors. This survey resulted in the detection and or- trail echoes. The head echoes are from a point-like bital characterisation of 109 major and minor meteor source, moving with the meteoroids and relatively in- showers, 12 of which were previously unreported dependent of aspect angle. Trail echoes come from sources. The EISCAT_3D radar will have the potential the ionised trail, and the conventional low-power me- to measure much more accurate orbital parameters teor radars (15–60 MHz) are able to detect only me- for micrometeors, because its sensitivity will be of a teor trails through specular reflection. In the early different magnitude compared with the CMOR radar, 1990s, the EISCAT UHF became the first ISR system making EISCAT_3D capable of measuring the orbits to be deliberately applied to the study of meteor head of smaller micrometeors at higher accuracy and pro- echoes (Pellinen-Wannberg and Wannberg 1994, viding new insight into a population that has not pre- Wannberg et al. 1996). Soon thereafter, all the world’s viously been measured continuously with good leading ISR facilities joined in. statistics. It is thus probable that the multi-static me- The present EISCAT has offered a possibility to study teor head echo observations that will be conducted the properties of the micrometeor population that can- by EISCAT_3D will result in the detection of numer- not be resolved by any other radar system. The most im- ous new sources of meteors and thus provide valuable portant of these has been the tristatic geometry of the new insight into the formation, distribution, and evo- UHF radar, which has made it possible to derive the full lution of dust within our solar system. velocity vector of the radar targets with very high accur- A recent study by Chau and Galindo (2008) showed acy. Observations have contributed to the database of that it is possible to distinguish parent sources for McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 32 of 63

Fig. 22 Meteoroid orbits calculated from tristatic EISCAT UHF measurements. Sun (yellow), Earth (blue), and prograde (green) and retrograde (red) meteoroid orbits (Szasz 2008) meteor head echoes measured with high-power large- et al. 2004). Such observations would improve the mass aperture radar. This shows that it is feasible to utilise flux estimates and also contribute to a better understand- such radars for mapping the distribution of micromete- ing of the physics governing the head echo process. oroids within our solar system. Furthermore, a recent Meteor head echo observations could be run in parallel extensive study by Kero et al. (2011) using the MU radar with standard ionospheric mode, which would produce a has shown that it is possible to measure meteor head semi-continuous meteor data set. echo trajectories in an automated fashion and hence to construct a large catalogue of meteor orbits. Planets and asteroids The total mass flux of meteors reaching the upper Planetary radar is a field of research which involves atmosphere is believed to deposit somewhere between using radars to study objects in our solar system. These 2 and 200 tonnes of material in Earth’s upper atmos- include the Sun, planets and their moons, comets, and phere daily (Murad and Williams 2002). The chemical asteroids (Gordon 1958; Ostro 1993). The advantage of composition and net mass of disintegrating meteors using a radar is the ability to control the signal that is are important, since they are related to many physical used to illuminate the target. This allows measurements and chemical phenomena such as polar mesospheric of various properties of targets through the use of time summer echoes and noctilucent clouds in the meso- delay, polarisation and Doppler shift. sphere as well as polar stratospheric clouds in the Planetary radar measurements can be used, among stratosphere. By applying simultaneous radar and op- other things, to determine and refine orbital elements tical imaging (multi-static video and spectral imagers), and spin vectors, to study surface and sub-surface com- small-scale fragmentation and sudden breakup, in position, and to study the shape and topography of which locked populations of volatile chemicals are as- planetary objects (Kaasalainen and Lamberg 2006; Ostro sumedtobereleasedfromthemeteoroidbody,could 1993). Most planetary radar work has been conducted be studied in great detail. Narrowband optical instru- with Earth-based radar systems, such as Arecibo, Gold- mentation using emission filters could help to identify stone, and the VLA, but recently, space probes have also chemical compounds such as water (cf. the “water- been used to conduct radar measurements of various Leonid” observations reported by Pellinen-Wannberg targets, such as the Moon (Bussey and Mini-RF 2010), McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 33 of 63

Venus (Pettengill et al. 1980; Saunders et al. 1990); and 2002). Measurements mainly conducted with the Are- Mars (Picardi et al. 2005). Ground-based planetary cibo Observatory radar (Ostro 1993) have provided new radar measurements typically involve measuring the information about several planets and a large number of same and opposite circular returns of the backscatter to asteroidal bodies. A high-power large-aperture radar is determine the surface reflectivity and roughness (Ostro also a valuable tool in accurately measuring the orbital 1993; Thompson 1978; Campbell et al. 2007). Because parameters of potentially hazardous near-Earth objects. the targets are typically far away, the spatial resolution Radar measurements can also provide information about is obtained by combining rotational Doppler shift and the shape and composition of these objects. time of delay. The resulting range-Doppler or delay- The main advantage of the EISCAT_3D system will be Doppler images of the targets are not completely unam- its complementary location compared with the US ra- biguous, as several different parts of a rotating object dars. This will allow opportunistic measurements of can result in identical Doppler shift and round-trip near-Earth objects that happen to pass Earth on our side delay. of the planet, i.e. if the US planetary radars cannot meas- EISCAT_3D can also be used for imaging solar system ure a near-Earth object passing Earth at closest ap- objects. Radar imaging of such objects is not the same proach, the EISCAT_3D system will be able to do this as optical imaging, since radar signals penetrate inside with a large probability. Radar measurements provide the object surface. The existing EISCAT UHF radar has complementary information about the range and non- been applied to range-Doppler mapping of the Moon convex shape of these objects, which can be compared (Fig. 23), and with the current facility, resolution down with optical measurements (Kaasalainen and Lamberg to about 60 m was obtained (Vierinen and Lehtinen 2006). An example of such a situation was the close ap- 2009). Polarimetric radar studies of the Moon are useful proach of 2012 DA14, which passed Earth with a dis- as they provide a way of probing the sub-surface geo- tance of only 26,000 km in January 2013. It was not chemical properties and the rock abundance of lunar possible to measure this asteroid with the US planetary regolith. Due to the lack of erosion, the Moon is thought radars at closest approach, but the existing set of EIS- to contain important clues to the formation of the CAT radars, located in northern Fennoscandia, secured Earth-Moon system and the statistics of meteoroidal and some good observations. asteroidal impacts on Earth. Lunar regolith is also a vi- The EISCAT_3D radar will also use a complementary able source for the 3He isotope, which is one possible fu- frequency compared with the US planetary radars. sion reactor fuel. Attempts have also been made to The Arecibo Observatory radar, which can use either image asteroids, but the present EISCAT system does 450 MHz or 3 GHz, is at much higher frequency not have high-enough resolution. EISCAT_3D will have than the 233-MHz frequency planned for the EIS- much improved timing accuracy, potentially allowing the CAT_3D radar. While the lower frequency results in technique to be extended to other planets and near- a smaller sensitivity for measuring near-Earth aster- Earth objects. oids, it provides larger ground penetration and thus Planetary radar measurements conducted with high- allows better sub-surface probing capabilities, allow- power large-aperture radars have resulted in a wealth of ing an improved measurement of the composition of new information about solar system bodies (Campbell planetary targets. The downside of the high-latitude location of EISCAT is that many of the planetary sources are at low elevation (20°–30°), significantly restricting the available sources that can be measured. This also poses a conflicting re- quirement for antenna sensitivity compared with iono- spheric plasma incoherent scatter radar measurements. As mentioned in “Structures and boundaries in the iono- sphere”, the EISCAT_3D design is a compromise, recog- nising that most EISCAT_3D measurements will be made at higher elevations, but choosing an antenna and array design which preserves a reasonable antenna gain down to elevations below 30° (Johansson et al. 2014).

Interplanetary scintillation from the solar wind The amplitude of a radio signal from a compact astro- Fig. 23 Lunar image produced by the EISCAT UHF radar (Vierinen nomical radio source, passing through the inner helio- and Lehtinen 2009) sphere, is modulated by the motion of solar wind plasma McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 34 of 63

irregularities across the line of sight. This modulation, Space weather and service applications which may be observed with a suitable radio telescope, Background is termed interplanetary scintillation (IPS). It is produced Over the last decade, a vibrant international community by variations in solar wind density (and thus the refract- has grown up around the study of space weather, focus- ive index) and subsequent interference. The turbulent- ing on the study of how varying conditions in geospace scale density irregularities in solar wind can be used as affect human activity (e.g. Baker 1998, 2002; Echer et al. flow tracers—assuming that their velocity is closely re- 2005; Bothmer and Daglis 2007). Space weather has a lated to that of the background solar wind (e.g. Hewish wide variety of impacts. For example, solar-terrestrial 1989). Simultaneous measurements by two antennas disturbances can produce significant changes in the show similar patterns of scintillation, and the time lag density of the ionosphere, particularly in the high- for maximum cross-correlation gives an estimate for the latitude regions, while electric fields can cause structur- solar wind outflow speed. Increasing the baseline be- ing of the ionospheric density into irregularities on a tween the antennas helps to resolve the presence of mul- wide range of spatial and temporal scales. These density tiple solar wind streams which can be located in the structures have serious impacts on radio propagation, same line of sight. As the baseline rotates relative to the both for trans-ionospheric signals from global position- solar wind outflow, estimates of the direction of flow ing satellites (e.g. Mitchell et al. 2005) and for ground- can be made. The existing Kiruna and Sodankylä UHF to-ground or ground-to-air communications reflected receivers have been used to carry out a series of world- from the ionosphere (e.g. Wilink et al. 1999). Processes leading observations of scintillating structures in the that heat the thermosphere-ionosphere system lead to solar wind showing, among other things, how rapidly the expansion of the neutral upper atmosphere, enhan- the solar wind is accelerated within a short distance after cing the thermospheric density at high altitudes and in- leaving the Sun, and how streams of different speeds can creasing satellite drag. Events such as coronal mass interact to produce complex solar wind structures (e.g. ejections (CMEs) and magnetospheric sub-storms in- Bisi et al. 2010; Dorrian et al. 2010). crease the flux of high-energy particles, which can dam- Since the solar wind structures (CMEs, CIRs) carry the age spacecraft electronics and increase the radiation enhanced fluxes of momentum and energy that drive exposure of astronauts, airline passengers, and air crew. geomagnetic activity, IPS measurements are potentially These phenomena hence have profound consequences an important space weather activity that can help to pre- for the multi-billion dollar aerospace and space technol- dict the onset of geomagnetic activity (Hapgood and ogy sector (e.g. Feynman and Gabriel 2000; Pirjola et al. Harrison 1994). The advantage of IPS measurements 2005). In addition, auroral ionospheric currents can over in situ measurements by satellites is their very good affect infrastructure on the ground, by inducing current spatial coverage, since the solar wind can be observed at flow in systems such as power grids and pipelines. There any latitude and observations can also cover a wide is thus strong interest from a variety of industrial sectors range of distances from the Sun. The temporal reso- in understanding and predicting space weather, so that lution is, however, often worse than in satellite measure- its effects can be mitigated. The ability to predict space ments. Use of several arrays and frequencies would give weather events requires highly capable models, able to improved coverage of the inner , reducing assimilate data from a diverse global network of con- seasonal variations in that coverage and offering a better tinuously observing instruments (e.g. Vassiliades 2000; possibility for cross-comparison to identify sources of Wehrenpfenning et al. 2001). Although incoherent interference and improve statistics. During the operating scatter radars, such as EISCAT_3D, are few in number, lifetime of EISCAT_3D, a number of other systems such the power and versatility of their measurement technique as LOFAR and SKA will be available, with the potential means that they can measure parameters which cannot be to make complementary solar wind observations, allow- obtained in any other way. EISCAT_3D will thus be one ing EISCAT_3D the possibility to play a role in a wide of the key cornerstones in the international endeavour to international coordination. measure and predict space weather effects. The data product needed for IPS is total power, inte- grated over a wide frequency band and sampled at >50 Hz. Space weather effects on high-latitude ionospheric The required receiver bandwidth must be equal to or irregularities larger than 20 MHz and contain minimal radio fre- A major effect of space weather on the ionosphere is the quency interference (RFI). In order to facilitate this sci- ability of solar-terrestrial processes to modulate and ence, EISCAT_3D should also have the ability to clip structure electron density. The term “ionospheric irregu- out individual narrow-frequency bands prior to integra- larities” indicates structures which are different from the tion over the bandwidth. The required elevation angles ambient ionosphere in electron density, either much are low (down to 5°), at all azimuths. higher or much lower. Their scale size can range from McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 35 of 63

centimetres to hundreds of kilometres. Irregularities can auroral zone. The lower panel shows the corresponding appear in a specific region or can be created over wide structures in electron temperature. The strong gradients areas by global-scale processes such as magnetic storms. at the edges of the patches are susceptible to the plasma They can occur on all timescales from seconds up to a instabilities which can create scintillation-producing ir- few days. While shorter timescale processes are always regularities (Kersley et al. 1988). As Fig. 25 shows, the present to some extent, they can be greatly enhanced existing EISCAT radars can only make a 1D cut through under specific space weather conditions. Space weather the highly structured ionosphere, whereas an imaging particularly affects the high-latitude regions, but can also radar such as EISCAT_3D will make full 3D images of act to produce density changes, waves, winds, and elec- the auroral ionosphere. tric fields on a global scale. In order to understand the When intense scintillation occurs, the integrity of formation of irregularities, it is necessary to measure all Global Navigation Satellite Systems (GNSS) such as of the processes responsible for their formation, includ- GPS, Glonass, and Galileo can be jeopardised. Under ing density gradients, particle precipitation, ExB drifts, extreme conditions, this corruption leads to the loss and thermospheric winds (Belehaki et al. 2009). of satellite lock and the total failure of global posi- Irregularities in the high-latitude region may result tioning, as occasionally happened during the last solar from patches of plasma density, large-scale plasma blobs, cycle (e.g. Basu et al. 2001; Webb and Allen 2004; and density troughs, whose steep edges are unstable, so Mitchell et al. 2005; Foster et al. 2005b). Since so that smaller scale density structures develop along these many aspects of the global economy now depend on edges. These smaller scale irregularities cause intense GNSS, the understanding of their vulnerabilities and scintillation effects. Individual patches of irregularities the development of risk-mitigating countermeasures have lifetimes of 2–3 h; however, irregularities are seen constitute key challenges (Fisher and Kunches 2011). to persist for periods of up to 8 h. These irregularities Current studies of irregularities frequently utilise the are not specifically related to space weather disturbances ground-based reception of GNSS signals to derive scintil- but do increase with the solar activity cycle. lation climatologies (e.g. Spogli et al. 2009; Alfonsi et al. At high latitudes within the auroral oval and cusp, 2011; Prikryl et al. 2011). In polar regions, however, the precipitating energetic particles produce enhanced density of GNSS receivers is sparse and supplementary electron densities. The fluxes of precipitating particles observations from other instruments are needed, not only are very structured in space and time and create ir- to improve the climatology but also to investigate the regular structures in the ionosphere. These types of physical processes behind irregularity formation. irregularities are very variable in space and time dur- Scintillation effects occur at all latitudes, but particu- ing space weather disturbances. larly at the equator as well as at high and polar latitudes To understand the importance of irregularities, con- (Aarons 1982; Mitchell et al. 2005), making northern sider the case of L-band scintillation. Figure 24 shows Scandinavia an ideal location for studying the formation schematically how scintillation is produced by iono- and evolution of the scintillation-producing irregularities spheric irregularities and how this affects satellite signals and the effects that they can have. In order to improve by refraction and diffraction. As radio waves propagate the characterisation of ionospheric irregularities, two through irregularities in the ionosphere, they experience types of EISCAT_3D observations will be needed. different values of TEC (total electron content), resulting Campaign-based observations will specify the real-time in group delay and phase advance, which are referred to distribution and short-term variations of irregularities in as refraction (e.g. Yeh and Liu 1982; Cerruti et al. 2008). the high-latitude region, while long-term monitoring is Diffraction arises when ionospheric irregularities form at essential to build climatologies of recurrent features in scale lengths of about 400 m and begin to scatter GPS their dynamics and temporal evolution (Alfonsi et al. signals. At the receiver, the GPS signals from different 2011). A complete understanding of irregularities can paths will add in a phase-wise sense, causing fluctuations only be achieved on the basis of continuous and system- in the signal amplitude and phase. On the ground, atic ionospheric monitoring, able to describe the tem- power fades may be deeper than 30 dB (Kintner et al. poral changes of the plasma on long timescales (months, 2009a, b). The relevance of EISCAT and EISCAT_3D ob- years, decades) as well as on short timescales (minutes, servations to these issues is extensively discussed in the hours, days) with a wide spatial coverage (hundreds of paper by Forte et al. (2013). kilometres). Such observations would be a superb The upper panel of Fig. 25 shows an example of large- complement to the monitoring data already available from scale ionospheric patches observed by the 32-m antenna other instruments (Altadil et al. 2009). Because of its abil- of the EISCAT Svalbard Radar looking to the south and ity to measure continuously over long periods and to observing plasma density patches as they pass over the image the ionosphere over wide spatial scales, EISCAT_3D polar caps and move equatorward towards the nightside will play a critical role in providing these measurements in McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 36 of 63

Fig. 24 Schematic showing scintillation effects in beacon satellite data, as produced by a highly structured ionosphere (image credit: University of Bath) the European high-latitude sector, thus providing new and can completely inhibit radio propagation, since signals insight into processes that lead to irregularity formation. can be completely absorbed in the lower ionosphere or Although we have focused so far on the creation of can propagate straight out into space without being ionospheric irregularities and their effect on satellite sig- reflected (e.g. Blagoveschensky and Borisova, 2000; Hun- nals, it should be noted that the same space weather sucker and Hargreaves 2002). As a result, there is a wide processes which produce strong density structures in the community of potential users needing reliable forecasts of high-latitude ionosphere also severely affect the propaga- the radio propagation environment in the high-latitude re- tion of radio signals at all frequencies from ELF to VHF, gion, where models are notoriously unreliable. The wide including commercial radio broadcasts and ground-to- coverage area and continuous operation of EISCAT_3D ground and ground-to-air communications. Reflections will make it an invaluable resource for users of the radio from sharp density gradients, such as auroral boundaries spectrum who need to constrain model predictions with or trough walls, can introduce large multi-path delays to real data (see “Space weather impacts on technology”). received signals, while fast plasma drifts in the ionosphere can give rise to spectral spreading and Doppler shifts in Space weather effects on the high-latitude radio signals. The most severe of all are the impacts ionosphere-thermosphere caused by geomagnetic storms and sub-storms, which can Space weather can also have very significant effects on the radically affect the density distribution of the ionosphere neutral atmosphere, especially at high latitudes. As noted

Fig. 25 EISCAT Svalbard Radar observations of southward-propagating ionospheric density patches and the underlying electron temperature structure (adapted from Bust and Crowley 2007) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 37 of 63

above, some of the most significant processes are associ- EISCAT data have already been used to constrain thermo- ated with geomagnetic storms, because such events can spheric density models at high latitudes (Bruinsma et al. last several days and have global effects. During storms 2003), and EISCAT_3D is expected to provide an even and sub-storms, large amounts of energy are injected into more suitable data set for this purpose, due to its consid- the polar upper atmosphere in the form of electric cur- erably enhanced temporal and spatial coverage. rents, accelerated particles, and Poynting flux. The sudden The ability of the ionosphere to carry electric currents heating of the upper atmosphere can change its compos- depends on the fact that the ionosphere is embedded in ition and chemistry, which in turn feed back into changes the neutral thermosphere. The differential motion of in the density and structure of the ionosphere. Some of ions and electrons under the action of magnetospheric the energy entering the polar atmosphere can also affect electric field gives rise to electric currents flowing hori- the neutral winds, changing the circulation of the upper zontally in the E region of the upper atmosphere. The atmosphere. In addition, storm and sub-storm effects gen- collisional interaction between ions and neutrals acts to erate a whole spectrum of atmospheric gravity waves, heat both species, a process which is conventionally which in turn produce travelling ionospheric disturbances referred to as Joule heating. Joule heating dissipates (TIDs). magnetospheric electromagnetic energy in the upper at- TIDs correspond to perturbations of the ionised gas mosphere. The effect of the heating is that the neutral associated with the passage of internal atmospheric upper atmosphere expands outward, so that the density gravity waves. They can give rise to horizontal gradi- of the high-altitude thermosphere increases during pe- ents in electron density, and, in this respect, their ef- riods of strong Joule heating. fects are similar to those of ionospheric irregularities. A significant effect of E region currents is that their A rich variety of publications has addressed the sub- rapid variations are able to cause induced currents in ject of TIDs (see review by Hocke and Schlegel conductive systems on the Earth’s surface, such as elec- 1996). The neutral wind and the Earth’s magnetic field tricity distribution grids and oil pipelines. Currents in- seem to be the main geophysical factors governing the duced by space weather effects can be a significant structure and dynamics of TID disturbances, although the problem for certain technological sectors, and these are physical mechanisms underpinning their relationship are discussed in “Space weather impacts on technology”. still not entirely understood (Afraimovich et al. 1999). The role of EISCAT_3D in investigating these phe- As well as the effect of thermospheric chemistry and nomena will be to provide underpinning data to help heat balance on the ionosphere, and the ability of ther- constrain and validate the models which are used in this mospheric waves to produce ionospheric structure area. We briefly discuss more of these models in “Mod- through TIDs, there are two other major space weather elling and forecasting of space weather”. EISCAT_3D effects involving the thermosphere, both of which rely will also play an important role in supporting current on the ability of space weather process to modulate the and future spacecraft missions designed to investigate electric currents which flow naturally in the ionosphere, the space weather effects arising from ionospheric cur- especially in the high-latitude region. rents. One such mission is ESA’s 3-satellite SWARM From a technological perspective, the first major mission, launched in November 2013 and with a nominal technological impact of the thermosphere is that modula- lifetime of 4 years, initially at altitudes of 460 and 530 km. tions in its density cause an increase or decrease in the at- The main objective of the SWARM mission is to provide mospheric drag on an orbiting satellite (e.g. Fedrizzi et al. the best ever survey of the geomagnetic field and its tem- 2012). Atmospheric drag has important implications not poral evolution. The geomagnetic field is made up of both only for operational spacecraft but also for space debris, to internal and external components. The internal field arises which we refer in “Space debris”. Despite the importance from the magnetism of the Earth’s core, the crust, and the of density modelling, however, this remains a challenging oceans. The external component arises from space wea- problem because of the number of factors influencing the ther processes, in particular from ionospheric currents. In neutral density, which include solar radiation, Joule heat- order to separate the internal and external contributions, ing, winds, and waves. A number of semi-empirical the three SWARM satellites carry instruments to obtain models are used by the spacecraft community, including unique new data of unprecedented resolution and accur- MSIS-86 (Hedin 1987), DTM-94 (Berger et al. 1998), acy for the global ionospheric science community. DTM-2000 (Bruinsma et al. 2003), and JB2006 (Bowman The SWARM instrumentation allows the estimation of et al. 2008); however, their accuracies are no better than electric currents and fields, ion and electron densities, and 10–15 % (Bruinsma et al. 2004), with the highest latitudes temperatures. From these, further parameters like the being the most unreliable. Figure 26 shows a comparison electromagnetic power (Poynting flux), electron cooling between four commonly used models on a latitude-time rates, etc. can be derived. The satellites’ acceleration indi- grid for a quiet day, revealing some notable differences. cates neutral density as well as cross-track winds. McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 38 of 63

Fig. 26 Neutral density at 425 km, represented on a latitude-time grid for day 200, a mean solar flux of 180 sfu, and Kp of 1, as predicted by the DTM-94 (a), MSIS-86 (b), DTM-2000 (c), and DTM-STAR (d) models (from Bruinsma et al. 2004)

Predecessors of the SWARM satellites, particularly it has been difficult to separate spatial from temporal vari- CHAMP, Dynamics Explorer, and the DMSP series, have, ations and to assign the proper scale sizes to these. In the in the past few years, already established the existence of near future, this will be possible, thanks to the volumetric relatively small-scale but strong density variations in the imaging capability of EISCAT_3D and the multi-satellite thermosphere (Lühr et al. 2004) and enhancements of the nature of the SWARM mission. Poynting flux at high latitudes (Knipp et al. 2011). There is currently a growing interest worldwide in the The causes of these variations are not yet well under- development of fleets of nano-satellites or “cubesats” stood. However, it is clear that they originate with the (size 10 × 10 × 10 cm). As an example, the QB50 net- space plasma processes which heat and force the upper at- work (led by von Karman Institute, Belgium) will com- mosphere. In terms of absolute magnitude, this takes pre- prise a set of 50 cubesats built by university teams and dominantly place in the lower thermosphere and E region, research institutes worldwide that will be launched to- well below the altitudes where satellites can orbit. These gether in early 2016 into a circular orbit at 320 km alti- height regions are, however, well covered by incoherent tude with an inclination of 79°. The duration of the scatter radars. Thus, only a combination of SWARM sat- mission is expected to be between 6 weeks and 3 months ellites and EISCAT_3D radar can deliver the measure- depending on solar activity. The objectives of QB50 are ments that allow us to understand magnetosphere- to explore in situ the low thermosphere and to study the ionosphere-thermosphere coupling more fully. Previously, atmospheric re-entry process. McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 39 of 63

Half of the cubesats of the QB50 mission will have a (Boteler et al. 1998). Extreme GIC events can cause multi-needle Langmuir probe (mNLP), developed at the problems especially to high-voltage power transformers, University of Oslo (Bekkeng et al. 2010). The instrument is which can even lead to blackouts. Significant GICs occur able to measure small-scale (down to tens of meters) dens- most typically during three different types of space wea- ity variations that result from the complex magnetosphere- ther events: auroral sub-storms, short-period pulsations, ionosphere interaction. For a good understanding of the and global sudden impulses (Viljanen et al. 2001; Viljanen physical processes, supplementary data are needed. These and Tanskanen, 2011). Sub-storms are thought to be the include simultaneous measurements from other spacecraft dominant process at high latitudes, due to the rapid onset (both in low and high Earth orbits) on roughly the same of ionospheric currents at the beginning of a sub-storm field line, the EISCAT_3D radar data, or ALIS data (iono- and the intensification of the large-scale electrojets during spheric optical tomography). QB50 Langmuir probe data its later phases (Viljanen et al. 2006). The large-scale and EISCAT_3D data could be used for cross-validation at current variations occurring in sub-storms are often 320 km. EISCAT_3D will also provide the overall iono- mixed with phenomena of smaller spatial scales, such as spheric context in which the mNLP data should be inter- vortex-like structures. preted. The goal would be to try to identify in what The forecasting of GIC is increasingly recognised as an situations density variations could be expected at the sat- important priority for Europe, and improved forecasting ellite altitude. techniques will be demonstrated within the FP7 EURIS- GIC project, during the period 2011–2014 (Viljanen Space weather impacts on technology 2011). The US solar shield method (Pulkkinen et al. 2010) In “Space weather effects on high-latitude ionospheric ir- is already in use in North America, and its output will be regularities”, we already looked at the impacts of space extended to the European region. The GUMICS-4 simula- weather on radio propagation and global positioning sys- tion of the Finnish Meteorological Institute (Janhunen tems. In this section, we look at geomagnetically induced et al. 2012) will also be enhanced to make it suitable for currents, already mentioned in “Space weather effects on forecasting purposes. the high-latitude ionosphere-thermosphere”, and at the The role of EISCAT_3D in GIC research would be to effects of energetic particles. provide the data needed for a comprehensive analysis of Geomagnetically induced currents (GICs) are related ionospheric electrodynamics based on solar wind- to rapid variations of the geomagnetic field (see Fig. 27) magnetosphere-ionosphere simulations. Although the and are observed in different technological conductor primary driver of GIC events is the solar wind, the condi- systems such as power grids and oil and gas pipelines tions in the ionosphere play a decisive role. From the

Fig. 27 Observed GICs at the Rauma 400-kV transformer and in the Nurmijärvi-Loviisa 400-kV power line, both in Finland, and the magnetic field as measured at Nurmijärvi on February 26, 1992 (Viljanen et al. 1999) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 40 of 63

viewpoint of GIC, understanding the temporal and spatial to energetic particle studies by providing measurements (3D) behaviour of ionospheric currents on scales of sec- to validate ionisation models for lower energy precipita- onds and a few tens of kilometres would be extremely tionintothemiddleatmosphere,whereEISCATobser- valuable. The combination of ground magnetic field re- vations are possible. In addition to the validation of cordings and ionospheric electric field recordings by EIS- models, climatologies of energetic particle data, derived CAT_3D will provide the tools needed to reach this goal. from EISCAT_3D measurements, will be examined to The Earth’s radiation environment comprises energetic assess whether they can be extrapolated to obtain useful particles trapped in the radiation belts and magneto- predictions e.g. to help facilitate the planning of safe air- sphere, those generated by solar energetic particle (SEP) craft routes. events, and galactic cosmic rays (GCRs). Of these differ- ent populations, the most highly variable, and conse- Modelling and forecasting of space weather quently the most difficult to anticipate, is the SEP “Ionospheric modelling” has already discussed some population. This is also the population that can often of the models used to describe the state of the cause the most damaging effects (Zank 2012). At cruise magnetosphere-ionosphere/thermosphere system, and altitudes of commercial aviation, both SEPs and GCRs given some details of how EISCAT_3D can be used to pose threats through single-event upsets in critical elec- verify and constrain such models. It should be noted tronics and through exposure of crew and passengers to here that the space weather community makes use of a radiation (Dyer et al. 2003; Barnard et al. 2011). The specific set of models, devoted to the prediction of iono- gravity of the consequences for biological structures spheric and geospace conditions. In general, these depends on the energy of the particles responsible models tend to be empirical or semi-empirical in nature, (Bottollier-Depois et al. 2003). Radiation effects are espe- since purely physics-based models are of limited use in cially significant for polar flights, because the cosmic radi- operational settings (e.g. Mikhailov et al. 2007). For ex- ation particle flux increases with increasing latitude and ample, the radio propagation community makes consid- altitude, being significantly higher on board aircraft than erable use of empirical models such as the global IRI, or at ground level; however, the complexity of the radiation the European COST-PRIME and COST-PROF family of field does not make dose measurements easy. Indeed, the models, which describe the average conditions of the un- particles encountered vary considerably, and a wide range perturbed, sub-auroral ionosphere (Bilitza 1992; Bradley of energies and types of particle are found. 1995; Hanbaba 1995; Zolesi and Cander 2004). Radiation standards for air crew are based on past ex- Forecasting of ionospheric conditions, e.g. for the perience, and there are no regulations concerning pas- coming 24 h, is of considerable importance for a variety senger exposure. Allowed radiation limits on the ground of communication, positioning, and frequency manage- vary between nations. The International Commission on ment tasks, especially during disturbed space weather Radiological Protection (ICRP) recommends a 20-mSv conditions. However, accurate forecasting remains a limit for the annual exposure of occupational radiation challenging problem, due to the wide variety of potential workers and a 1-mSv annual limit for the public and forcing mechanisms and the diversity of their possible prenatal exposure (Wrixon 2008; Mertens et al. 2010). effects. One of the fundamental problems is the lack of Dosages during a flight depend on the path, duration, data on the statistical variability of, and the interrelation- and altitude as well as on the level of solar activity. For ship between, different effects such as sporadic E, spread example, Mertens et al. (2010) found that a commer- F, blobs and patches, irregularity formation, and changes cial 8‐h polar flight during the 2003 “Halloween” SEP in the height profile of electron density. The construc- event would have given 0.7 mSv. Models such as tion of better forecast models remains the subject of a QARM (Lei et al. 2006) show that a round trip of broad international effort, but this relies on strong two such flights during the recent solar minimum underpinning science driven by more complete and bet- gave a GCR dose of order 0.2 mSv, while an equivalent ter coordinated sets of long-term observational data. trip during the largest known SEP event, the “Carrington Even the best physics-based models are capable of event” of 1859 (Shea et al. 2006; Hapgood 2011), would making relatively basic errors, as shown when they are have given 20 mSv. Estimates vary, but from studies of confronted with real data. Figure 28 shows a month- smaller events, the direct and knock‐on (loss of service, long time series of continuous field-aligned measure- etc.) costs of a Carrington‐scale event have been estimated ments made by the EISCAT Svalbard Radar in April at $1–2 trillion and full recovery would take 4–10 years 2007 as part of the International Polar Year. The red (Odenwald et al. 2006). horizontal line shows the average altitude of the F region Although EISCAT_3D is unlikely to measure directly peak density during the period, as derived from the mea- the most energetic particles which penetrate into the surements, while the blue horizontal line shows the peak lower atmosphere, EISCAT_3D data will still contribute altitude predicted by the CMAT model from University McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 41 of 63

Fig. 28 ESR measurement of electron density (Ne) altitude profiles during April 2007, with the average position of the measured Ne peak shown in red and the position of the predicted peak indicated in blue. Note the large disparity between the measurement and the model predictions (courtesy of A.P. van Eyken)

College London. A major discrepancy of some 60 km is 2006). In addition, during very disturbed conditions, the noted between modelling and measurement. A working signal is absorbed in the ionosphere. Observing tech- group hosted by the International Space Science Insti- niques including incoherent scatter radars (Farley 1996) tute has been devoted to studying the comparison of like EISCAT_3D, sounding data from satellites (Reinisch data and models during the IPY period and studying the et al. 2001b), ionospheric tomography (Leitinger 1996), or reasons for the differences encountered. radio occultation measurements (Jakowski 2005) can be While it is not impossible that pure physics-based used to obtain the missing information. modelling could eventually provide reliable space wea- Both modelling and real-time specification are particu- ther predictions, given increased understanding and de- larly challenging at high latitudes, where the ionosphere velopment, a better approach for operational use in the tends to be considerably more structured. In principle, short term is in the development of assimilative models this implies the need to assimilate a greater density of (e.g. Schunk et al. 2004; McNamara et al. 2007; Angling observations; however, the distribution of ground-based 2008), in which real-time data are used to describe glo- observing instruments is often the most sparse at the bal conditions and to constrain models which seek to highest latitudes, putting a greater premium on the as- predict the future evolution of the system over a rela- similation of satellite data, e.g. from Langmuir probes tively short period. In this respect, the situation is simi- and radio occultation techniques, or data derived from lar to meteorology—while we cannot predict the GPS tomography. weather from first principles, we can forecast quite suc- One of the primary goals of the space weather com- cessfully for a limited period by accurately measuring munity is to secure an improved set of real-time and his- the current conditions and using physical models to ex- torical data on which to base specification and empirical trapolate the short-term evolution of the weather and modelling. This implies not only the deployment of climate system. more observing instruments but also better international For the determination of current space weather condi- collaboration to maximise data availability and improve tions and their impact on the ionosphere, real-time spe- data exchange (Stanislawska and Belehaki 2009). A var- cification is usually preferable to modelling. The iety of organisations have responsibility for the manage- problem with specification, however, is that it relies on ment of data relevant to space weather, including ESA, the assimilation of real-time observations which are EUMETSAT, Intermagnet, ISES, COST, DIAS, and the often very sparse, meaning that underlying theoretical World Data Centres. The present EISCAT data, as well constraints are needed, for example to regularise the ob- as the future EISCAT_3D data, will be highly relevant to servation grid. Certain types of observation themselves this challenge, implying a requirement for better need to be supplemented by modelling. For example, methods of disseminating and assimilating EISCAT data. one of the most widely available observing techniques is Since 2011, EISCAT has been actively involved in two ground-based vertical-incidence HF sounding; but this European initiatives leading towards the development of technique cannot provide any information on the shape such a framework. The first such initiative is ESPAS, a of the topside ionospheric profile, and models based on consortium of European Space Weather data providers, the ionospheric-scale height are used to provide this infor- spanning radars, scintillation studies, magnetic field mation (e.g. Reinisch et al. 2001a; Belehaki and Kersley measurements, and a wide range of other techniques McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 42 of 63

(see Fig. 29). Among other targets, the ESPAS study will users of these products would include not just the space provide a consistent data format which allows all of weather and atmospheric science communities but also these different data types to be used interactively. The the education sector and the general public. other such initiative is ENVRI, formed by a consortium The wide-scale coverage and continuous operational of all the ESFRI environmental facilities, concerned with capability of EISCAT_3D will make it a superb tool for developing common approaches to data curation, distri- specification of the high-latitude European sector and bution, and sharing between the next generation of for the provision of data to initialise and constrain fore- European environmental projects, including a number of cast models, with the potential ability to provide detailed test cases based on real geophysical systems. EISCAT is real-time ionospheric maps of a region some hundreds a leading player in ESPAS and will have a smaller but of thousands of cubic kilometres in extent. EISCAT_3D still influential role in ENVRI. ENVRI-like initiatives are is complementary to other instruments, since while also being funded in physical sciences, biology, and so- tomographic techniques, for example, can provide a cial sciences, and EISCAT_3D is keeping closely abreast wide-scale context, EISCAT_3D can provide the de- of all these developments. In summary, the field of data tailed multi-parameter data to help explain why a par- interoperability for European scientific infrastructures ticular phenomenon is happening, not just where it seems poised for an upward step in capability. occurs. Collaborations have already been established Many of the concepts described above are captured in with the leading ionospheric tomography project in the a recent proposal to develop a common framework cov- Scandinavian sector (TOMOSCAND, Vierinen et al. ering all the main types of observing systems for moni- 2010), and the practical details of how the interaction toring near-Earth space, together with the current between the two projects will proceed are already under generation of models. The development of such a sys- active consideration. tem, known as the “Geospace Array” (Lind 2011), has re- As well as providing the basic underpinning data cently been summarised in a white paper to NASA. It needed to generate and validate space weather models, calls for a combination of globally distributed small- EISCAT_3D will have the capability to reconfigure its scale and medium-scale instruments, supported by operations to respond to changing space weather condi- facility-class instruments such as EISCAT_3D, all of tions. The use of “intelligent scheduling” (see “Observing which contribute data to a “Geospace Assimilation Grid” techniques and measurement philosophies”) will enable which in turn provides data to a range of search engines, the radar to initiate special experiment modes in re- virtual observatories, and space weather models. The sponse to particular space weather events, such as solar

Fig. 29 Instruments and data providers participating in the ESPAS consortium (image credit: ESPAS, http://www.espas-fp7.eu/trac/wiki/ PublicPages/ESPASConsortium) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 43 of 63

flares, magnetic impulses, or sudden storm commence- Iridium-Cosmos satellite collision (Vierinen et al. 2009a, b). ments. This capability will overcome a fundamental limi- Figure 30 shows a beam-park measurement of the tation of the present EISCAT radars which, because they Iridium-Cosmos collision debris conducted using the are not operated continuously, often fail to observe the EISCAT UHF radar. very dynamic conditions which occur at the start of such One of the most effective and cost-effective active mea- events, whose effects are of particular interest for the sures that can be taken is the so-called orbital collision space weather community. avoidance manoeuvre. In order to reduce the number of these manoeuvres and minimise the amount of fuel re- Space debris quired to perform them, accurate orbital elements of Europe has recently begun making efforts to establish a centimetre-scale orbital debris are needed. EISCAT is cur- space situational awareness (SSA) programme, under the rently building a capability for measuring orbital elements aegis of the . This programme is of space objects, with several successful campaigns con- designed to provide an independent capability for moni- ducted in the recent years in collaboration with ESA. toring spacecraft and the growing amount of “space deb- However, the slowly moving, single-beam nature of the ris”, ranging from large objects, such as dead satellites, current EISCAT radars makes it difficult to track space- to the millions of sub-centimetre fragments in Earth craft or debris objects. orbit (Klinkrad and Jehn 1992; Klinkrad 2006). While The design of EISCAT_3D enables it to overcome space debris poses no immediate danger to most many of the limitations inherent in monitoring space humans, there is a severe risk to satellites and space mis- debris with dish-based radars. Its wide spatial coverage sions operating in low Earth orbit, which provide many and capability to generate multiple, rapidly moving useful services to society. beams will enable EISCAT_3D to track individual ob- Currently, there are approximately a million objects jects, including multiple objects simultaneously, for an larger than 1 cm in orbit. Any of these objects would be optimal characterisation of their orbital parameters and likely to cause serious damage in the case of a collision the monitoring of orbit perturbations. The EISCAT_3D with an operational satellite. The situation is made worse radar will therefore be able to provide the high-accuracy by the so-called Kessler syndrome (Kessler and Cour- measurements required for orbit determination, and it is Palais 1978), a modelling result that predicts that the probable that such measurements can be “piggybacked” current orbital debris will collide with other orbital ob- onto normal incoherent scatter measurements, with jects, creating ever more small debris and further in- relatively sparsely spaced dedicated tracking pulses creasing the risk of orbital fragmentation events. This aimed at further refining the measurement of orbital ele- collisional cascade could potentially render many of the ments. Satellite tracking and space debris measurements low Earth orbital regimes unusable in the next 100 years in EISCAT_3D will also benefit from some of the other if no active measures are taken. new system capabilities, including the use of aperture Space debris monitoring has three main goals: firstly, synthesis imaging techniques to improve the resolution to identify and monitor the most significant objects and of sub-beam width-sized hard targets. debris clouds and characterise their orbits; secondly, to Figure 31 shows the smallest perfectly conducting validate the current models of how the space debris spheres detectable using the existing EISCAT radars and population is evolving over time, for example monitoring EISCAT_3D. The calculation assumes that the scattering the dispersal rates of debris clouds and quantifying the is either Rayleigh or optical, depending on the size. tendency of debris populations to cascade into progres- Without significant loss of applicability, the Mie scatter- sively smaller fragments (Rossi et al. 1998); and thirdly, ing, occurring between these two size scales, is ignored, to measure how the locations and orbital parameters of as it is a second-order correction that applies for objects the space debris population are modified by specific with dimensions comparable with the radar wavelength. space weather events, such as geomagnetic storms, We assume that a signal to noise ratio of 10 dB after co- which can perturb their motions or cause them to de- herent integration is needed for reliable detection of a orbit through increases in thermospheric drag. target. As the specifications for EISCAT_3D, we have as- The existing EISCAT radars are sensitive to small deb- sumed a 25-% duty cycle, 230-MHz frequency, 2-MW ris fragments in the centimetre size range, and EISCAT peak power, and 150-K receiver noise temperature. We has been involved in statistical studies of orbital debris have assumed that the antenna is either four or eight for over 10 years (Markkanen et al. 2002, 2005, 2009), by times larger than the current VHF antenna, which re- conducting so called beam-park measurements of debris. sults in antenna directivities of 52 and 55 dB (personal These measurements have shed light on the two major communication by J. Vierinen 2012). fragmentation events that have occurred in the recent The sensitivity of the proposed EISCAT_3D system years, the Chinese anti-satellite experiment and the in both cases is comparable to the existing UHF McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 44 of 63

Fig. 30 Comparison between measurements of the space debris environment as observed by the EISCAT UHF radar (left) and the predictions of the ESA PROOF model (right) (Vierinen et al. 2009b). The clouds of vertically distributed points correspond to debris fragments from the Cosmos-Iridium collision radar. Based on these assumptions, the EISCAT_3D reduce the risk of collisions between space objects system will be capable of detecting objects with 1.5– and orbiting spacecraft. 2-cm diameter at a distance of 1000 km. There are The high-latitude location of the EISCAT_3D radar currently very few space surveillance radar systems has pros and cons. The biggest disadvantage is that the capable of routinely tracking such small-sized objects, radar will not be able to measure objects with low incli- although new capabilities are being built around the nations, as these will never fly over the radar. However, world. EISCAT_3D would thus be a very valuable most of the orbital debris is in high-inclination Sun- asset for the global space surveillance network, pro- synchronous orbits, which means that it will pass over viding orbital element measurements of small but po- EISCAT_3D. Due to this fact, the density of orbital deb- tentially harmful objects, which could be used to ris is higher at high latitudes, meaning that EISCAT_3D

Fig. 31 Space debris diameter vs. range for EISCAT_3D and the existing EISCAT radars (personal communication, Vierinen 2012) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 45 of 63

will be able to observe more debris per unit volume than conventional incoherent scatter radar applications but also radars located in more southern locations. This gives for carrying out passive radar applications over multiple EISCAT_3D the potential to conduct orbital parameter regions of frequency space and undertaking generic moni- surveys of space objects at a faster rate than radars at toring functions for applications such as space weather more southerly locations. and space situational awareness. Phased arrays (such as EISCAT_3D) offer many add- Radar techniques and new methods for coding itional degrees of freedom compared to conventional and analysis dish-based radars. Among others, these include the pos- Background sibilities for modularising the radar array, for imaging Over the 30-year period for which the EISCAT facilities applications (Grydeland et al. 2005b); post-beam forming have been in operation, one of their key contributions to (e.g. Hansen 2009), for example to produce multiple re- the field of upper atmospheric science has been in the ceiver beams within the volume of a single transmitter development of new observing techniques to improve beam; aperture tapering to achieve finely tuned control the quality, time resolution, and spatial resolution of the of the beam shape; and array coding to generate twisted radar measurements. EISCAT has always been a test bed beams (Tamburini et al. 2012). The necessity to develop for new ideas in coding and data analysis, whose user an entirely new transmitter system for EISCAT_3D community has pioneered many novel applications, in- makes it possible to include further additional capabil- cluding new coding strategies (e.g. Virtanen et al. 2009; ities, such as the possibility to handle amplitude modula- Lehtinen et al. 1998, 2008), new types of data analysis tion and higher resolution phase and polarisation (e.g. Lehtinen and Huuskonen 1996; Virtanen et al. 2008), control, which are not practical in the current EISCAT and other applications of fundamental mathematical and radars. statistical theory to experiment design and data taking In order to achieve these aims, considerable amounts (e.g. Vallinkoski and Lehtinen 1991). Many of these new of computing power will need to be devoted to the sig- techniques, first developed at EISCAT, are now in stand- nal processing and analysis of the EISCAT_3D data ard use among incoherent scatter radars worldwide. (Mann et al. 2013). Because the low-level EISCAT_3D EISCAT_3D represents a further substantial step in the data will be very large, some of this capability will need design of atmospheric radars. The system will be the first to be deployed at the radar sites and will be devoted to of a new generation of software radars (e.g. Grydeland data decimation, data combination, short-term storage, et al. 2005a), whose advanced capabilities will be rea- and event detection. Many of these functions will have lised not by its hardware (which is relatively inexpensive to be automated, since the volume of data is so large as and modular) but by the flexibility and adaptability of to be beyond the ability of a human operator to interpret the scheduling, beam forming, signal processing, and them in real time. Because of this, it is also desirable to analysis software used to control the radar and process store data at the lowest level possible for a limited its data. In this respect, EISCAT_3D will be a world period, so that decisions made automatically can be re- leader in the development of new observing techniques, vised in the light of later reconsideration, for example if which will eventually be implemented by the next gen- interesting events are found to develop from initially un- eration of incoherent scatter radars around the world. promising conditions. The use of EISCAT_3D to test innovative radar tech- The remainder of the EISCAT_3D storage and com- niques aimed at major advances in resolution, measure- puting capabilities will be provided off-site, utilising the ment speed, coverage, occupancy, consistency, and power of supercomputing and archiving centres in the responsiveness is needed not just to increase the scien- various participating countries. These centres will pro- tific output of the facility itself but to position the whole vide the Web-based remote control, data visualisation, international radar community to move forward over and experiment design capabilities required for the oper- the coming years. ation of EISCAT_3D, as well as the intelligent technol- EISCAT_3D is designed to provide a high-power, ogy that will make it possible for the radar to respond high-gain aperture, combining fully digital element tech- adaptively to changing conditions, e.g. by changing its nology with a wide receiver bandwidth over a large aperture characteristics, look directions, coding schemes, number of simultaneous, adaptive, and independent re- and data recording rate. ceiver beams. The system will be able to carry out large- The following sub-sections discuss some specific appli- scale volumetric imaging of the upper atmosphere, while cations which, while impossible with the current gener- also providing high-resolution images to probe small-scale ation of EISCAT radars, will form a fundamental part of targets beyond the reach of existing radars. In addition, the observing capabilities of EISCAT_3D. In each of these EISCAT_3D will have the ability to conduct a number applications, EISCAT_3D will serve as a “pathfinder” in- of different experiments simultaneously, not only for strument, where new techniques will be developed into McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 46 of 63

standard observing applications for implementation in the trade-offs between coverage area, observation density, next generation of software radars. and time resolution in constructing a volumetric image. An interesting aspect of this is that the optimum im- Observing techniques and measurement philosophies aging strategy is highly condition dependent, since the Because EISCAT_3D is a very different type of radar to scattering cross-section is dependent on the density of the existing EISCAT systems, and has considerably more the ionosphere and is likely to be significantly more vari- flexibility, it opens the possibility for a radical re- able than might be the case for weather radar. Another appraisal of the principles under which such radars issue will be the existence of ionospheric structure at should be operated in future. The major novel tech- various different scales, depending on time, location, and niques opened up by the new facility are as follows: magnetic activity level, which determines the density at which the volumetric image should be sampled. This Volumetric Imaging The ability of phased array radars means that the optimum imaging strategy will need to to change their beam direction almost instantaneously be varied in real time during experiments as conditions gives EISCAT_3D considerable potential to act as a change, and strategies for doing this appropriately will volumetric radar, producing images of the type shown need to be evolved. In certain circumstances, there will schematically in Fig. 32. The volumetric imaging carried be contradictory demands, for example if the ionosphere out by EISCAT_3D will not necessarily be genuinely has low density and a high degree of structure, such that simultaneous, since the coverage is necessarily limited potentially no volumetric imaging strategy will be able by the number and scan speed of the transmitter beams, to adequately represent it. The chosen approach to volu- and there will always be a requirement to trade the metric imaging is a complex problem, which not only in- number of directions being scanned against the accuracy volves a trade-off between scan speed and measuring and time resolution of the resulting image. Nonetheless, volume but also involves trade-offs between parameters the imaging capacity of EISCAT_3D will be considerably such as transmitter power and beam shape (see below). superior to that of any existing incoherent scatter system The task of developing an automated approach to the and will enable quasi-simultaneous wide-field imaging of handling of volumetric imaging, balancing time integra- the ionosphere to be undertaken for the first time. This tion and spatial resolution in an optimal way, will be technique is not completely new, since some similar im- crucial for the next generation of imaging upper atmos- aging philosophies are already in use. One example is pheric radars and one in which EISCAT_3D will have a the VIPIR (Volumetric Imaging and Processing of Inte- path-finding role to play (Wannberg et al. 2010). grated Radar) technique used in the weather radar com- Post-acquisition beam forming is a critical technique munity, for which high-order processing and graphics on the receiver side of the system, needed to ensure that capabilities have already been developed. EISCAT_3D the volumes illuminated by the radar transmitter are be- will seek to exploit these developments while pioneering ing appropriately measured. In the context of a multi- their application in the upper atmospheric context. element array, the challenge is mainly in the computa- One of the most important requirements for this ap- tion and data handling capacity required to ensure that plication is to find an intelligent way to balance the enough receiver beams are generated to map the full

Fig. 32 Schematic representation showing the volumetric imaging of an ionospheric layer above Scandinavia by EISCAT_3D (image courtesy of EISCAT Scientific Association) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 47 of 63

extent of each transmitter beam, with required range resolution. As with volumetric imaging, changes in condi- tions will imply changes in the required resolution, espe- cially if the ionosphere changes from a relatively quiet state to one having a high degree of horizontal and vertical structure. This type of change is very likely at the auroral latitudes where EISCAT_3D will operate. Another circum- stance requiring an adaptive approach to beam forming would be the need to study a particular region of interest as it propagates through an observing region illuminated by a known transmitter pattern. It is worth re-iterating that, in contrast to the changes in volumetric imaging Fig. 33 Prof. Cesar La Hoz (University of Tromsø) standing beside philosophy discussed above, these changes in beam form- one of the panel antenna arrays currently being used for testing the ing strategy can be achieved simply by data post- aperture synthesis imaging concept at the EISCAT Svalbard Radar, processing. The challenge is equivalent to an inversion prior to its implementation in EISCAT_3D problem in which the measured set of all samples from every radar antenna is inverted in an optimal way to re- possibility would exist to reconfigure the experiment to cover the large-scale structure of the scattering ionosphere, study such structures in more detail, either by changing subject to certain constraints on temporal and spatial reso- the illumination strategy or by altering the post-processing lution imposed by the measurement strategy. EISCAT users to optimise the imaging of a particular region. The avail- already have considerable experience with the application able computing budget (in terms of operations per second) of inverse problems to radar experimentation, and the use will probably be a key constraint here, which has to be of inversion techniques to produce optimal estimates of traded against the storage capacity for low-level data, since wide-scale ionospheric structure through beam forming is it is unlikely that the optimum inversion can be carried out likely to represent a significant contribution to the future inrealtime,especiallyinthecaseofahighlystructured generation of ionospheric phased array radars, which is well ionosphere. The methods adopted to arbitrate between matched to the community’s existing capabilities. these different constraints, and the ways by which they are applied under varying conditions, will again be a significant Aperture synthesis imaging radar The same kinds of contribution from EISCAT_3D to inform the operation of inversion and image recovery techniques are equally ap- a new generation of radar systems. plicable to the imaging of small-scale, sub-beam width structures, as well as to the large-scale morphology of Beam and phase front shaping A further degree of the ionosphere. The aperture synthesis imaging tech- freedom is introduced into phased array radar experi- nique is already being used at the EISCAT Svalbard ments by the ability to control the beam shape and the Radar, employing a small number of additional low-gain shape of the phase fronts being generated. This tech- arrays distributed around the main radar site, an ex- nique has many potential applications, some of which ample of which is shown in Fig. 33 (see, for example, are intimately related to the imaging and inversion tech- Schlatter et al. 2013). The existence of a distributed re- niques described above. A simple example of beam ceiving array, as planned for EISCAT_3D, with receiving shaping is the possibility to vary the width of the radar elements located up to 2 km from the array centre, al- beam in order to distribute the transmitter power over a lows the reconstruction of ionospheric targets down to wider area (e.g. Kinsey 1997). This has clear application tens of metre scales, significantly smaller than the beam to imaging, since scanning a wider beam can cover the widths of the current generation of incoherent scatter sky more quickly than scanning a narrow beam. The radars (La Hoz and Belyey 2011). As before, the most downside, of course, is that a broader beam implies a significant part of the imaging challenge lies in data lower power per unit volume and also a coarser reso- post-processing, since the ensemble of samples from the lution. Since the beam forming used on the transmitter distributed array elements will already exist but will need side need not be the same as that on the receiver side, to be inverted to recover the small-scale structure. One however, the possibility exists to use a very wide trans- of the key aspects of this technique, therefore, consists mitter beam to simultaneously illuminate a large region of determining when it is actually needed, in other of ionosphere, the structure of which can then be words determining when (and where within the viewing resolved by forming multiple narrow receiver beams, as area) any small-scale structures exist. A second key as- in Fig. 34 (Lind et al. 2002). This would ameliorate the pect is to determine how the system should respond second of the above problems. Depending on the charac- when a small-scale structure is identified, since the teristics of the system, and the width of the beam, the McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 48 of 63

Fig. 34 Representation of the “holographic radar” concept, in which multiple distributed passive sites are used to spatially resolve the volume illuminated by a transmitted beam, via the formation of multiple simultaneous received beams (Lind et al. 2002)

first problem might nonetheless mean that such a tech- At these altitudes, the phase fronts emanating from a dish nique may only be applicable to an ionosphere above a antenna of equivalent size to the EISCAT_3D array would certain density threshold, since the power per unit volume be highly non-planar, making it complicated to interpret could otherwise be low enough for the backscatter from a the instantaneous signal received from a broad, distributed, weak ionosphere to fall below the detectability threshold horizontally stratified target such as the atmosphere. As of the system. The choice of when to use a narrow beam, long as the beam shape and the physics of the scattering and when to use a broad beam, is therefore a complex process are known, of course, any such scattering problem one, related to the aims of the experiment being under- can be handled by inversion techniques, but the ability to taken and the changing conditions. In situations charac- conduct adaptive beam shaping to reduce the degree of terised by spatially structured and very rapidly changing non-planarity in the radar phase fronts can simplify the processes, however, genuinely simultaneous spatial cover- problem by reducing the need for computationally inten- age provided by multiple receiver beams in a broad illumi- sive inversion. The ability to control phase front shape may nated volume would be preferable to an imaging strategy thus be important in the plans to use EISCAT_3D for me- based on the rapid scanning of transmitter beams. teorological applications. Although tropospheric and lower The possibility to use shaped phase fronts will be of par- stratospheric studies are not the main purpose of the radar, ticular interest when using the large EISCAT_3D aperture this is expected to be a scientifically interesting application in the near field, i.e. for lower atmosphere measurements. which opens the radar up to a wider user community. McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 49 of 63

Event identification and object tracking One of the measurement (Dobrinsky and del Monte 2010). This type significant improvements which EISCAT_3D will bring of mode has potential to add significantly to the scientific over the existing EISCAT radars is the ability to track return of the new radar and will develop a technique moving objects, which has been impossible with the which will be widely used in other radars of this type. current generation of large, slowly scanning dishes. Such objects can be natural, such as plasma blobs, troughs, Adaptive interleaved experiments It is implicit in many flow channels, and aurora, which need to be followed so of the above sections that EISCAT_3D will radically re- that their evolution can be observed, or man-made ob- define the experimental philosophy of incoherent scatter jects, such as space debris, which need to be tracked so radars. In contrast to the current situation, where the radar that their orbital parameters can be calculated. runs a single experiment at a time, the new generation of Although EISCAT_3D will have the capability to track phased array radars will be able to share their duty cycle objects such as satellites on the basis of known orbital pa- between a number of different modes, whose operations rameters, a key technique for the new radar will be to will be interleaved. For example, although the radar might track objects whose motion is not known in advance. This devote a substantial proportion of its duty cycle to a pre- will require the capacity to identify objects and respond to defined mode of ionospheric observation, a fraction of its them in real time, through an algorithm which predicts observing time could be simultaneously devoted to other their future motion based on the previous few seconds applications such as space debris, while some of its receiver of observation, and changes the beam positions of the beams could be used for radio astronomy observations, e.g. radar to respond accordingly. Figure 35 shows how this for solar wind studies via interplanetary scintillation. can be done, using a grid of beams whose positions and A major issue, which will drive the development of spacing can be varied in real time. These capabilities will techniques for all phased array radars of this type, is the enable EISCAT_3D to function in an event-driven mode, question of how these interleaved experiments should responding rapidly as geophysical conditions change, change as conditions alter. This question has many as- and also to contribute effectively to space situational pects: not only how the experimental modes themselves awareness studies such as space debris acquisition and should change, in terms of beam direction, modulation

Fig. 35 Schematic showing how the motion of a space debris object, indicated by the red line, could be tracked by using multiple quasi- simultaneous beams, as will be available in EISCAT_3D (courtesy of Dr J Vierinen) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 50 of 63

scheme, or transmitted power, but also how the division of the facilities has been the development of new methods of observing time should be re-assigned between com- of coding and data analysis to optimise the scientific return peting applications, how the computing resources should from the radars. These include the operational development be re-distributed between different kinds of data pro- of techniques such as the flexible re-use of radar samples in cessing, and, perhaps most importantly, the identifica- correlation (the GEN system, Turunen 1986), the use of tion of what circumstances would trigger such changes modulations based on Walsh codes (alternating codes, e.g. in the observational philosophy. Häggström et al. 1989, Wannberg 1993, Lehtinen et al. EISCAT_3D will have a capacity to serve scientists and 1998), and analysis based on the principles of statistical in- science in a better way than the present EISCAT has by en- verse theory (GUSDAP, Huuskonen and Lehtinen 1996; abling a more flexible radar scheduling system. While it will Lehtinen and Huuskonen 1996). Many of these techniques still be possible to assign time in advance to pre-defined are in use, not only at EISCAT but also elsewhere in the modes, there will be considerable emphasis on a condition- worldwide network of incoherent scatter radars. dependent response to changing conditions. This requires The development of new techniques has, however, now the development of a strong set of criteria specifying what reached the point at which some of the newly designed ex- conditions would trigger a change of mode, or what class of periments cannot be tested operationally at EISCAT, given phenomenon would require an interruption of “normal” the limitations of the current EISCAT hardware, though observations, or a significant re-distribution of resources, in they have been tested elsewhere (e.g. Virtanen et al. 2013). order to make detailed observations of its development. For example, the work on polyphase codes, exploring how One method of achieving this, which should be pioneered additional degrees of freedom can be opened up by the using EISCAT_3D, is to make a very strong connection be- control of the transmitted polarisation to levels of polar- tween data and models, with radar (and other) data being isation change less than 180°, cannot be exploited because assimilated to models in real time and used to predict the the current EISCAT transmitters are not capable of the re- future development of the geophysical processes which, in quired phase control. In addition, the use of amplitude turn, will trigger new modes of observation. This will need modulation in EISCAT coding, which has been extensively considerable development effort, since both the modelling studied by the Finnish EISCAT user community, cannot and the assimilation infrastructure are not currently at the be implemented in practice because the present transmit- level where they can fully support such a system, and there ter exciters are not capable of the required amplitude con- is clear potential for wrong decisions to be made, if the pre- trol. These limitations will not exist in EISCAT_3D, for dictions are not good enough to be up to the challenge. which the transmitter is being designed not only to be suf- The whole field of adaptive experiments and intelligent ficiently flexible to transmit all of the coding strategies scheduling is a significant area of study in itself, with which have so far been devised by the EISCAT commu- considerations potentially extending beyond pure sci- nity, but also to accommodate the next round of coding ence into issues such as logistics and economics. It en- developments anticipated during the coming years. capsulates the challenge of how to derive the maximum The preparations for EISCAT_3D are already driving benefit from a multi-faceted resource, which not only some new developments in the types of experiments avail- can be represented in terms of transmitter and comput- able at EISCAT. Some of the recently developed coding ing power but also needs to be evaluated in terms of its strategies, such as sidelobe-free “perfect codes” (Lehtinen scientific and technical benefit. et al. 2009), codes using ramped interpulse periods, or with Many of these issues, which are fundamental to the oper- modulation patterns based on arithmetic progressions (see ationofEISCAT_3D,arestudiedintheHandbook of Meas- Fig. 36), can already be adapted for use on the existing urement Principles which is being produced as an output of hardware, though enhanced computing capabilities are Work Package 6 on Performance Specification (Lehtinen needed for the optimum analysis of these data. These will et al. 2014). The handbook makes a first-principles introduc- be taken into use in the standard operation of the existing tion to the fundamental underpinnings of phased array ob- radars during the coming months, as part of the develop- servations, and discusses the various trade-offs which apply ment undertaken in Work Package 11 of the Preparatory in the case of each measurement strategy. These will form Phase (Software Development) with new computers being the basis for a first definition of the operating modes of the deployed to the Tromsø site to handle the required data new radar, and will feature heavily in the succeeding genera- processing and storage. tions of the Performance Specification Documents, which While there are certain new coding applications which will be the other major outputs from Work Package 6. cannot be tested on the existing EISCAT system because of the limitations of the current hardware, there is another Radar coding and data analysis methods class of experiments which can only be run on the EIS- One of the most prominent contributions which the EIS- CAT_3D system because they are designed specifically for CAT community has made over the whole 30-year operation phased arrays. These experiments use the principle of McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 51 of 63

Fig. 36 Examples of radar codes based on the use of ramped interpulse periods, or arithmetic progressions in transmission time, which are currently being evaluated for use in EISCAT_3D experiments (image courtesy of Juha Vierinen)

“antenna coding”, in which the shape of the array is used indications that this can be the case, for example in to add an extra degree of freedom to the transmitted “plasma vortices” formed as a result of ionospheric modifi- modulation. An example, which has been the subject of cation experiments (Leyser et al. 2009, see “Ionospheric considerable recent study in the optics and geophysics modelling”). communities, is the use of twisted beams (see Fig. 37), Although probing for ionospheric signatures of orbital generated by applying a progressive phase shift to the angular momentum is not the main justification for EIS- transmission in order to generate a beam which carries or- CAT_3D, the technology can in principle be deployed bital angular momentum (Leyser and Wong 2009). The easily on the new radar, and the novelty value of an ac- conservation of orbital angular momentum determines tual demonstrated detection makes it well worthwhile to that, if there are any structures in the upper atmosphere plan for this kind of experiment to be included in the which have angular momentum signatures, these signa- repertoire of EISCAT_3D. tures would be detectable from the phase of the returned signal. This principle has been used in optics to study the Conclusions properties of photons, since orbital angular momentum This paper is a revised version of Deliverable 3.6 of the (OAM) signatures in photons can arise as a consequence EISCAT_3D Preparatory Phase Study, funded by the of variations of phase and intensity in optical fields. European Union as part of its Seventh Framework Programme. No practical demonstration has yet been made of The project deliverable version can be found online at the whether OAM information can be extracted from upper EISCAT_3D website, at https://www.eiscat3d.se/content/ atmosphere experiments, but there are some theoretical deliverable-36-final-version-eiscat3d-science-case.

Fig. 37 Comparison of standard and twisted beams (left and right respectively), showing the type of experiment potentially needed to probe orbital angular momentum structures in the upper atmosphere (Leyser et al. 2009) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 52 of 63

The intent of the science case document has been to For the science topics outlined in the science case, the set out a broad scope of achievable science aims, with- three science working groups discussed the kinds of obser- out going into great detail on any one topic in particu- vations that EISCAT_3D should make. In addition, a lar. In doing so, it offers a basis for users in each number of experts have been contacted for input. The con- potential EISCAT_3D member country to accentuate clusions are summarised in Tables 1 and 2. For each topic, particular science areas for their own national funding the key plasma parameters to be measured are identified in applications, which should be geared to their own re- Table 1, together with the required time resolution and search strengths. altitude extent, and the optimal horizontal and vertical resolutions for measurements in the E region, unless other- Appendix wise stated. Also noted in Table 2 are the required horizon- EISCAT_3D measurements and complementary instruments tal coverage, optimal location of the radar, the ideal A1. Measurement requirements for the various science number of quasi-simultaneous beams, other requirements, topics and useful complementary ground-based instruments.

Table 1 EISCAT_3D resolution and range extent requirements for the different science topics. A phased array system with fast pointing, multiple beams and calibrated signal is assumed Science topic Parameter for which Temporal Horizontal resolution Vertical Height range resolution is given resolution (s) (km) resolution (km) (km)

Mesoscale electrodynamics and flows Ne, Te, Ti, Vi 10 20 in the F region 2 85–400 (including BBFs)

Small-scale (auroral) dynamics Ne 1 1 0.5 70–500

Te, Ti, Vi –“– –“– –“– 85–400

Fine-scale auroral structures Ne, Te, Ti 0.1 0.1 0.2 70–200

Vi 0.1 0.1 5 120–400

Ion outflow (natural and heater-induced) Ne, Te, Ti, Vi, ion comp. 10 10 20 200–1500 NEIALs (aperture synthesis imaging) Raw data 0.03 0.05 at 300 km 1 100–1500

Ionospheric irregularities Ne, Te, Ti, Vi 11 190–800 + + + Topside composition (O ,He ,H ) mi (and Ne, Te, Ti, Vi) 10 N/A N/A 300–1500

Transition region composition mi (and Ne, Te, Ti, Vi) 10 N/A 10 km 100–300 + + + (NO /O2 vs. O )

High-energy particle events (SEPs, etc.) Ne 110150–400

Te, Ti, Vi –“– –“– –“– 100–400

Atmosphere-ionosphere coupling Ne, Te, Ti, Vi, Vn <1 min 1 0.1 or better As low as (AGW, winds) possible—120 km

Mesosphere-stratosphere-troposphere Vector neutral wind, Ne <1 min 1 0.1 or better As low as (MST) small-scale dynamics possible—110 km

D region phenomena Ne, Te (=Ti) Vi (=Vn) 1 1 0.3 70–90 PMSE, PMWE Raw data, Doppler <1 min 1 0.1 or better 55–95 velocity, spectral width Meteoroids and their effects on the Raw data, polarisation 1 ms 0.01 0.01 (30) 70–200 background (Es, PMSE etc.), high-power matrix, and Ne, Te, Ti, Vi (1000) mode Planets and asteroids Raw data, power, 10-MHz sampling 15 m polarisation matrix Interplanetary scintillation Raw data 0.01 N/A N/A N/A

Heating experiments Ne, Te, Ti, Vi 1 1 1 100–2000 Heating experiments, aperture synthesis Spectra (raw data) IPP (~2 ms) 0.01–0.05 0.1 100–300 imaging Space debris monitoring and satellite Raw data, power, 10-MHz sampling 15 m tracking Doppler velocity Meteoroid monitoring Raw data, polarisation IPP ~100 ms for 0.01 Down to (30) 70–200 (piggyback and low-power mode) matrix, and Ne, Te, Ti, Vi low-power mode 10 m (1000) McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 53 of 63

Table 2 EISCAT_3D radar coverage and location requirements for different science topics, together with suggestions for complementary instruments Science topic Horizontal coverage Optimal location of radar Number of beams; other Complementary ground-based requirements instruments Mesoscale electrodynamics As wide as possible (30° Within auroral oval 20 × 20; multi-static Magnetometers, optical and flows (including BBFs) elevation with all azimuths) meas. equipment, tomography radio receivers, SuperDARN Small-scale (auroral) dynamics 100 km in the F region, Within the auroral oval 30 in N-S, 10 in E-W; Optical equipment, rockets 33 km in the E region Multi-static meas. – “– – “– – “– – “– Fine-scale auroral structures 5 km in the E region, within Within the auroral oval 5–7 N-S, 1–7 E-W; multi- Optical equipment, rockets 10–20 km from magnetic static meas., interfero- zenith in the F region metric studies – “– – “– – “– – “– Ion outflow (natural and ~100 km curvature of field Within the auroral oval N-S fan of 5–10 beams Optical equipment, rockets heater-induced) line NEIALs (aperture synthesis 40 km × 40 km at 700 km Within the auroral oval Spreading the beam in Optical equipment, rockets imaging) the F-A direction; plasma lines ±6 MHz Ionospheric irregularities As wide as possible Within the auroral oval 20 × 20 GPS scintillation measurements, rockets Topside composition N/A (field line tracing) Fan, enough to cover Optical equipment, rockets (O+,He+,H+) field line curving Transition region composition N/A One; heater ion- Optical equipment, rockets + + + (NO /O2 vs. O ) cyclotron freq. gives additional information High-energy particle events As wide as possible Oval and just 20 × 20 VLF, riometers, MF radar, (SEPs, etc.) equatorward of the oval satellites (POES, GOES, LANL, etc.), rockets –“– –“– Atmosphere-ionosphere As wide as possible Auroral oval and at the 20 × 20 MST radar, MF radar, meteor coupling (AGW, winds) edge of the polar vortex radar, lidar, airglow imager, FPI, sounding rockets, heater (API) Mesosphere-stratosphere- As wide as possible Location inside the polar 20 × 20 (M), one (ST); MST radar, MF radar, meteor troposphere (MST) small-scale winter vortex favourable additional low-power TX radar, lidar, airglow imager, dynamics for some questions mode, dual frequency sounding rockets, heater (200–1000-kHz separation) for interferometric distance determination D region phenomena As wide as possible Oval and just 20 × 20 Riometer, VLF, MF radar, rockets, equatorward of the oval heater PMSE, PMWE As wide as possible 20 × 20 MST radar, additional ISR at higher frequency, lidar, sounding rockets, heater Meteoroids and their effects Fattened beam or multiple Tracking mode (20 × 20 Meteor radar, MST radar, optical on the background (Es, PMSE, beams piggyback), fan for networks etc.), high-power mode receiver beams, multi-static measurements for trajectory Planets and asteroids Low elevation look As far south as possible One; arbitrary Other IS radars, optical direction to the south, at polarisation TX and RX, instruments least 25° hydrogen maser clocks Interplanetary scintillation Low elevation with all One; BW > 20 MHz, Other radio receiver networks, azimuths, particularly to the minimal RFI and ability including LOFAR and e-MERLIN south to clip out individual narrow-frequency bands prior to integration over the bandwidth McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 54 of 63

Table 2 EISCAT_3D radar coverage and location requirements for different science topics, together with suggestions for complementary instruments (Continued) Heating experiments 20° zenith angle Auroral oval 10 × 10 MST radar, airglow imager, FPI, sounding rockets Heating experiments, 20° zenith angle Fatten the beam or MST radar, airglow imager, FPI, aperture synthesis imaging multiple beams (10 × 10) sounding rockets Space debris monitoring 25° elevation with all North for satellite N/A; multi-static meas. and satellite tracking azimuths tracking for trajectory, multiple overlapping receiver beams for interferomet- ric angle determination, dual freq. (200–1000-kHz separation) for interfero- metric distance determination Meteoroid monitoring N/A N/A Meteor radar, MST radar, (piggyback and low-power optical networks mode)

Competing interests Lucilla Alfonsi, INGV Rome, Italy, [email protected] As stated elsewhere in this document, the work of preparing the EISCAT_3D Hervé Lamy, Belgian Institute for Space Aeronomy, Belgium, science case was supported by the Seventh Framework programme of the [email protected] European Union, as part of the EISCAT_3D Preparatory Phase Study. Dr. McCrea Frederic Pitout, Observatoire Midi Pyrénées, France, [email protected] and Dr. Aikio were both part funded to coordinate this action. In addition, Iwona Stanislawska, Space Research Centre, Poland, [email protected] members of the science working groups (see below) received EU travel support Juha Vierinen, Sodankylä Geophysical Observatory, Finland, [email protected] to attend working group meetings. Dr. McCrea also received travel support In addition, Stephan Buchert ([email protected]) and Thomas Leyser ([email protected]), from from the Japanese Geophysical Union to present the findings at its 2014 the Swedish Institute of Space Physics in Uppsala, took part in the work of meeting in Yokohama, on condition that a summary would be prepared for this group, and we have included them as coauthors. publication in Progress in Earth and Planetary Science (this paper). Group Three 2012/13 (radar techniques and complementary instruments) Björn Gustavsson, University of Tromsø, , [email protected] Authors’ contributions Craig Heinselman, EISCAT Scientific Association, [email protected] As convenors of the EISCAT_3D Science Working Group, AA and IM chaired Johan Kero, Swedish Institute of Space Physics, Kiruna, , all of the working group meetings which took place during the EISCAT_3D [email protected] Preparatory Phase project. The members of the working groups in each of Mike Kosch, University of Lancaster, UK, [email protected] the first three years of the study were as shown in the Authors’ Information section, above. In general, each working group member wrote a section of Acknowledgements text relating to their specialist area; IM and AA then integrated these The authors wish to recognise the contributions of the European Union’s contributions into the science case paper, with IM drafting and distributing Seventh Framework programme, which supported the meetings of the the final version. All authors read and approved the final manuscript. working groups and funded the time of the convenors to construct the science case document. We would also like to thank the staff of the EISCAT Authors’ information Scientific Association for their assistance and support. In particular, we The EISCAT_3D science case was constructed by three working groups, express our appreciation to Stephan Buchert and his colleagues at IRF drawn from the international EISCAT user community, with a rotating Uppsala, for hosting the annual series of EISCAT_3D User Meetings, which membership. In each of the first 3 years of the study, the working group had provided the venue for several of our working group discussions and were a a specific theme, meaning that the majority of the group were specialists in rich source of ideas and discussions. In developing the science case, we have that area. In the first year, the group specialised in atmospheric science, sought to reach out to researchers from both within and outside the current while the second year concentrated on space physics and space weather, member countries of EISCAT. As well as participants from Norway, Sweden, and the third year focused on radar techniques and complementary Finland, , and the UK, we are pleased to acknowledge the instrumentation. The convenors, Drs. Aikio and McCrea, were members of contributions of our colleagues from Belgium, France, , Italy, and each group and helped to facilitate the discussions. A preliminary version of Poland. Their experience and expertise have helped considerably to the science case was produced at the end of each year, becoming strengthen this document. progressively more detailed and complete as the study evolved. The author affiliations for all of the participants were given at the beginning of the Author details paper. The membership of each group, and their emails, were as follows: 1RAL Space, STFC Rutherford Appleton Laboratory, Harwell, Oxfordshire, UK. approved the final manuscript.Working Group Convenors 2Department of Physics, University of Oulu, Oulu, Finland. 3Istituto Nazionale Anita Aikio, University of Oulu, Finland, [email protected] di Geofisica e Vulcanologia (INGV), Rome, Italy. 4Swedish Institute of Space Ian McCrea, STFC Rutherford Appleton Laboratory, UK, [email protected] Physics (IRF), Kiruna, Sweden. 5Swedish Institute of Space Physics (IRF), Group One 2010/11 (specialising in atmospheric science) Uppsala, Sweden. 6British Antarctic Survey, High Cross, Cambridge, UK. Mark Clilverd, British Antarctic Survey, UK [email protected] 7Leibniz Institute of Atmospheric Physics, Kuhlungsborn, Rostock, Germany. Norbert Engler, IAP Kuhlungsborn, Germany, [email protected] 8Department of Physics and Technology, University of Tromsø, Tromsø, Yasunobu Ogawa, National Institute for Polar Research, Japan, Norway. 9EISCAT Scientific Association, Kiruna, Sweden. 10Department of [email protected] Physics, University of Lancaster, Lancaster, UK. 11South African National Space Kjellmar Oksavik, University of Bergen, Norway, [email protected] Agency, Hermanus, South Africa. 12Belgian Institute for Space Aeronomy, Markus Rapp, German Space Agency, Germany, [email protected] Brussels, Belgium. 13National Institute of Polar Research, Tachikawa, Tokyo, In addition, Evgenia Belova ([email protected]), from the Swedish Institute of Japan. 14Birkeland Centre for Space Science, University of Bergen, Bergen, Space Physics in Kiruna, made valuable written and oral contributions to the Norway. 15Department of Physics, University of Umeå, Umeå, Sweden. 16IRAP, work of this group, and we have included her here as a coauthor. UPS-OMP, Université de Toulouse, Toulouse, France. 17IRAP, CNRS, 9 Avenue Group Two 2011/12 (space physics and space weather) du Colonel Roche, Toulouse, France. 18Institute of Atmospheric Physics, McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 55 of 63

German Aerospace Center, Oberpfaffenhofen, Wessling, Germany. 19Space Barnard L, Lockwood M, Hapgood MA, Owens MJ, Davis CJ, Steinhilber F (2011) Research Centre, Warsaw, Poland. 20Sodankylä Geophysical Observatory, Predicting change. Geophys Res Lett 38:L16103. doi:10.1029/ Sodankylä, Finland. 2011GL048489 Basu S et al (2001) Ionospheric effects of major magnetic storms during the Received: 28 October 2014 Accepted: 3 July 2015 International Space Weather Period of September and October 1999: GPS observations, VHF/UHF scintillations, and in situ density structures at middle and equatorial latitudes. J Geophys Res 106:30389–30413. doi:10.1029/2001JA001116 References Baumjohann W, Paschmann G, Lühr H (1990) Characteristics of high-speed ion Aarons J (1982) Global morphology of ionospheric scintillations. Proc IEEE flows in the plasma sheet. J Geophys Res 95:3801–3809 – 70:360 378 Becker E, Schmitz G (2002) Energy deposition and turbulent dissipation owing to Afraimovich EL, Boitman ON, Zhovty EI, Kalikhman AD, Pirog TG (1999) Dynamics gravity waves in the mesosphere. J Atmos Sci 59:54–68. doi:10.1175/1520-0469 and anisotropy of travelling ionospheric disturbances as deduced from Bekkeng TA, Jacobsen KS, Bekkeng JK, Pedersen A, Lindem T, Lebreton JP, Moen transionospheric sounding data. Radio Sci 34:477–487 JI (2010) Design of a multi-needle Langmuir probe system. Meas Sci Technol Aikio AT, Selkälä A (2009) Statistical properties of Joule heating rate, electric 21(8):085903 field and conductances at high latitudes. Ann Geophys 27:2661–2673 Belahaki A, Kersley L (2006) Statistical validation of a technique for estimating Aikio AT, Opgenoorth HJ, Persson MAL, Kaila KU (1993) Ground-based total electron content from bottomside ionospheric profiles. Radio Sci measurements of an arc-associated electric field. J Atmos Terr Phys 41:RS5003. doi:10.1029/2005RS003433 55:797–808 Belehaki A, Stanislawska I, Lilensten J (2009) An overview of ionosphere- Aikio AT, Mursula K, Buchert S, Forme F, Amm O, Marklund G, Dunlop M, Fontaine D, thermosphere models available for space weather purposes. Space Sci Vaivads A, Fazakerley A (2004) Temporal evolution of two auroral arcs as Rev 147:271–313. doi:10.1007/s11214-009-9510-0 measured by the Cluster satellite and coordinated ground-based instruments. Belikovich VV, Benediktov EA, Tolmacheva AV, Bakhmet’eva NV (2002) Ionospheric Ann Geophys 22:4089–4101 research by means of artificial periodic irregularities. Copernicus GmbH, Germany Aikio AT, Pitkänen T, Kozlovsky A, Amm O (2006) Method to locate the polar cap Belova E, Chilson P, Kirkwood S, Rietveld MT (2003) A time constant of response boundary in the nightside ionosphere and application to a substorm event. of PMSE to ionospheric heating. J Geophys Res 108:8446 Ann Geophys 24:1905–1917. doi:10.5194/angeo-24-1905-2006 Belova E, Smirnova M, Rietveld MT, Isham B, Kirkwood S, Sergienko T (2008) Akasofu SI (1964) The development of the auroral substorm. Planet Space First observation of the overshoot effect for polar mesosphere winter Sci 12:273–282. doi:10.1016/0032-0633(64)90151-5 echoes during radiowave electron temperature modulation. Geophys Res Akhmedzhanov RA, Vikharev AL, Gorbachev AM, Ivanov OA, Kolganov NG, Lett 35:L03110. doi:10.1029/2007GL032457 Kolisko AL, Ofitserov MM (1995) Nanosecond microwave discharge as an ozone source in the upper atmosphere. Phys Lett A 207(3–4):209–213 Berger C, Biancale R, Ill M, Barlier F (1998) Improvement of the empirical Akmaev RA, Fomichev VI, Zhu X (2006) Impact of middle atmospheric thermospheric model DTM: DTM-94 comparative review on various temporal – composition changes on greenhouse cooling in the upper atmosphere. variations and prospects in space geodesy applications. J Geod 72:161 178 J Atmos Sol Terr Phys 68:1879–1889. doi:10.1016/j.jastp.2006.03.008 Biebricher A, Havnes O, Hartquist TW, La Hoz C (2006) On the influence of Alfonsi L, Spogli L, De Franceschi G, Romano V, Aquino M, Dodson A, Mitchell CN plasma absorption by dust on the PMSE overshoot effect. Adv Space Res – (2011) Bipolar climatology of GPS ionospheric scintillation at solar minimum. 38(11):2541 2550 Radio Sci 46:RS0D05. doi:10.1029/2010RS004571 Bilitza D (1992) Solar-terrestrial models and application software. Planet Space Sci – Altadil D, Boska J, Cander LR, Gulyaeva T, Reinisch BW, Romano V, Krankowski A, 40:541 544 – Bremer J, Belehaki A, Stanislawska I, Jakowski N, Scotto C (2009) Near Earth Bilitza D (2001) International reference ionosphere 2000. Radio Sci 36(2):261 276 ’ space plasma monitoring under COST 296. Ann Geophys 52:221–234 Bisi MM, Fallows RA, Breen AR, O Neill IJ (2010) Interplanetary scintillation Amm O, Viljanen A, Pulkkinen A, Sillanpää I, Vanhamäki H (2003) Methods for observations of stream interaction regions in the solar wind. Sol Phys – combined ground-based and space-based analysis of ionospheric current 261(1):149 172. doi:10.1007/s11207-009-9471-1 systems. Proc. Fourth Oersted International Science Team Conference, Blagoveshchensky DV, Borisova TD (2000) Substorm effects of ionosphere and HF – Copenhagen, Denmark, pp 181–184, Sep 23–27, 2002 propagation. Radio Sci 35(5):1165 1171 Amm O, Donovan EF, Frey H, Lester M, Nakamura R, Wild JA, Aikio AT, Dunlop M, Blelly PL, Robineau A, Lilensten J, Lummerzheim D (1995) 8-moment fluid models Kauristie K, Marchaudon A, McCrea IW, Opgenoorth HJ, Strømme A (2005) of the terrestrial high-latitude ionosphere between 100 and 3000 km. Solar Coordinated studies of the geospace environment using Cluster, satellite and Terrestrial Energy Program Ionospheric Model Handbook ground-based data: an interim review. Ann Geophys 23:2129–2170 Blelly PL, Lilensten J, Robineau A, Fontanari J, Alcaydé D (1996) Calibration of a Andrews DG, Holton JR, Leovy CB (1987) Middle atmosphere dynamics. numerical ionospheric model with EISCAT observations. Ann Geophys Academic, San Diego, Calif, p 489, ISBN: 0120585766/0-12-058576-6 14:1375–1390. doi:10.1007/s00585-996-1375-x Angelopoulos V, Kennel CF, Coroniti FV, Pellat R, Kivelson MG, Walker RJ, Russell Blelly PL, Lathuillère C, Emery B, Lilensten J, Fontanari J, Alcaydé D (2005) An CT, Baumjohann W, Feldman WC, Gosling JT (1994) Statistical characteristics extended TRANSCAR model including ionospheric convection: simulation of of bursty bulk flow events. J Geophys Res 99:21257–21280 EISCAT observations using inputs from AMIE. Ann Geophys 23:419–431 Angelopoulos V, McFadden JP, Larson D, Carlson CW, Mende SB, Frey H et al (2008) Blix TA (1999) Small scale plasma and charged aerosol variations and their Tail reconnection triggering substorm onset. Science 321(5891):931–935 importance for polar mesosphere summer echoes. Adv Space Res Angling MJ (2008) First assimilations of COSMIC radio occultation data into the 24(5):537–546 Electron Density Assimilative Model (EDAM). Ann Geophys 26:353–359 Blix TA, Thrane EV, Andreassen O (1990) In situ measurements of fine-scale struc- Archer D and Pierrehumbert R (eds.) (2011) The warming papers. Wiley ture and turbulence in the mesosphere and lower thermosphere by means Arnold NF, Robinson TR (1998) Solar-cycle changes to planetary wave propagation of electrostatic positive ion probes. J Geophys Res 95:5533–5548 and their influence on the middle atmosphere circulation. Ann Geophys Blixt EM, Grydeland T, Ivchenko N, Hagfors T, La Hoz C, Lanchester BS, Løvhaug 16:69–76 UP, Trondsen TS (2005) Dynamic rayed aurora and enhanced ion-acoustic Arnold NF, Robinson TR (2001) Solar magnetic flux influences on the radar echoes. Ann Geophys 23:3–11 dynamics of the winter middle atmosphere. Geophys Res Lett Borovsky JE (1993) Auroral arc thicknesses as predicted by various theories. 28(12):2381 J Geophys Res 98(A4):6101–6138 Austin J, Butchart N, Shine KP (1992) Possibility of an Arctic ozone hole in a Boteler DH, Pirjola RJ, Nevanlinna H (1998) The effects of geomagnetic disturbances doubled-CO2 climate. Nature 360:221–225 on electrical systems at the Earth’s surface. Adv Space Res 22:17–27 Baker DN (1998) What is space weather? Adv Space Res 22(1):7–16 Bothmer V, Daglis IA (2007) Space weather—physics and effects. Praxis Baker DN (2002) How to cope with space weather. Science 297(5586):1486–1487 Publishing, Chichester, UK, and Springer Science + Business Media, LLC, Baldwin MP, Dunkerton TJ (1999) Propagation of the Arctic Oscillation from the New York, NY USA. ISBN 978-3-540-23907-9 stratosphere to the troposphere. J Geophys Res 104:30937–30946 Bottollier-Depois JF, Chau Q, Bouisset P, Kerlau G, Plawinski L, Lebaron-Jacobs L Baldwin MP, Dunkerton TJ (2001) Stratospheric harbingers of anomalous weather (2003) Assessing exposure to cosmic radiation on board aircraft. Adv Space regimes. Science 294:581–584 Res 32(1):59–66 McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 56 of 63

Bowman BR, Tobiska WK, Marcos FA, Valladares C (2008) The JB2006 empirical Chilson PB, Belova E, Rietveld MT, Kirkwood S, Hoppe UP (2000) First artificially thermospheric density model. J Atmos Sol Terr Phys 70:774–793 induced modulation of PMSE using the EISCAT heating facility. Geophys Res Bradley PA (1995) Review of COST 238 (PRIME) achievements, IEE colloquium on Lett 27:3801–3084 HF antennas and propagation, 7/1-7/6 Cho JYN, Kelley MC (1993) Polar mesosphere summer radar echoes: observations Brasseur G, Solomon S (1986) Aeronomy of the middle atmosphere, 2nd edn. and current theories. Rev Geophys 31(3):243–265 Reidel, Norwell, Mass Cho JYN, Röttger J (1997) An updated review of polar mesosphere summer Brattli A, Blix TA, Lie-Svendsen Ø, Hoppe UP, Lübken FJ, Rapp M, Singer W, echoes’ observation, theory, and their relationship to noctilucent clouds and Latteck R, Friedrich M (2006) Rocket measurements of positive ions subvisible aerosols. J Geophys Res 102(D2):2001–2020 during polar mesosphere winter echo conditions. Atmos Chem Phys Clilverd MA, Seppälä A, Rodger CJ, Verronen PT, Thomson NR (2006) 6:5515–5524 Ionospheric evidence of thermosphere-to-stratosphere descent of polar Brekke A (1979) On the relative importance of Joule heating and the Lorentz NOx. Geophys Res Lett 33:L19811. doi:10.1029/2006GL026727 force in generating atmospheric gravity waves and infrasound waves in the Collis PN, Häggström I, Kaila K, Rietveld MT (1991) Eiscat radar observations of auroral electrojets. J Atmos Terr Phys 41(5):475–479 enhanced incoherent scatter spectra; their relation to red aurora and Bremer J, Hoffmann P, Latteck R, Singer W (2003) Seasonal and long-term field-aligned currents. Geophys Res Lett 18:1031–1034. doi:10.1029/91GL00848 variations of PMSE from VHF radar observations at Andenes, Norway. Crutzen PJ (1970) The influence of nitrogen oxides on the atmospheric ozone J Geophys Res 108:8438. doi:10.1029/2002JD002369 content. Q J R Meteorol Soc 96:320–325 Bremer J, Hoffmann P, Höffner J, Latteck R, Singer W, Zecha M, Zeller O Crutzen PJ, Isaksen ISA, Reid GC (1975) Solar proton events: stratospheric sources (2006) Long-term changes of mesospheric summer echoes at polar and of nitric oxide. Science 189:457–458 middle latitudes. J Atmos Sol Terr Phys 68(17):1940–1951 Czechowsky P, Reid IM, Rüster R, Schmidt G (1989) VHF radar echoes observed in Brewer AW (1949) Evidence for a world circulation provided by the the summer and winter polar mesosphere over Andøya, Norway. J Geophys measurements of helium and water vapor distribution in the stratosphere. Res 94(D4):5199–5217. doi:10.1029/JD094iD04p05199 Quart J Roy Meteor Soc 75:351–363 Dahlgren H, Aikio A, Kaila K, Ivchenko N, Lanchester BS, Whiter DK, Marklund GT Brown P, Weryk RJ, Wong DK, Jones J (2008) A meteoroid stream survey using (2010) Simultaneous observations of small multi-scale structures in an auroral the Canadian Meteor Orbit Radar I. Methodol Radiant Catalog, Icarus arc. J Atmos Sol Terr Phys 72:633–637 195:317–339 Deng Y, Ridley AJ (2007) Possible reasons for underestimating Joule heating Bruinsma S, Thullier G, Barlier F (2003) The DTM-2000 empirical thermospheric in global models: E field variability, spatial resolution, and vertical model with new data assimilation and constraints at the lower boundary: velocity. J Geophys Res 112:A09308. doi:10.1029/2006JA012006 accuracy and properties. J Atmos Sol Terr Phys 65:1053–1070 Diloy PY, Robineau A, Lilensten J, Blelly PL, Fontanari J (1996) A numerical model of Bruinsma S, Tamagan D, Biancale R (2004) Atmospheric densities derived from the ionosphere, including the E-region above EISCAT. Ann Geophys 14:191 CHAMP/STAR accelerometer observations. Planet Space Sci 52:297–312 Djuth FT, Sulzer MP, Elder JH, Wickwar VB (1997) High-resolution studies of Bussey B, Mini-RF Team (2010) Results from the first year of Mini-RF operations atmosphere-ionosphere coupling at Arecibo Observatory, Puerto Rico. on Lunar Reconnaissance Orbiter. AGU Fall Meeting Abstracts, A7 Radio Sci 32:2321–2344 Bust GS, Crowley G (2007) Tracking of polar cap ionospheric patches using data Djuth FT, Sulzer MP, Gonzales SA, Mathews JD, Elder JH, Walterschied RL (2004) A assimilation. J Geophys Res 112:A05307. doi:10.1029/2005JA011597 continuum of gravity waves in the Arecibo thermosphere. Geophys Res Lett. Callis LB, Lambeth JD (1998) NOy formed by precipitating electron events in doi:10.1029/2003GL019376 1991 and 1992: descent into the stratosphere as observed by ISAMS. Dobbin AL, Aylward AD, Harris MJ (2006) Three-dimensional GCM modelling of Geophys Res Lett 25:1875–1878 nitric oxide in the lower thermosphere. J Geophys Res 111(A):07314. Callis LB, Natarajan M (1986a) Ozone and nitrogen dioxide changes in the doi:10.1029/2005JA011543 stratosphere during 1979–84, 19. Nature 323:772–777. doi:10.1038/ Dobrinsky N, del Monte L (2010) The space situational awareness programme of 323772a0 the European Space Agency. Cosm Res 48(5):392–398 Callis LB, Natarajan M (1986b) The antarctic ozone minimum: relationship to Dobson GMB (1956) Origin and distribution of the polyatomic molecules in the odd nitrogen, odd chlorine, the final warming, and the 11-year solar atmosphere. Proc Roy Soc London 236A:187–193. doi:10.1029/2009JA014766 cycle. J Geophys Res 91:10771–10796. doi:10.1029/JD091iD10p10771 Donaldson JK, Wellman TJ, Oliver WL (2010) Long-term change in thermospheric Callis LB, Boughner RE, Baker DN, Mewaldt RA, Blake JB, Selesnick RS, Cummings temperature above Saint Santin. J Geophys Res 115 A11. doi: 10.1029/2010JA015346 JR, Natarajan M, Mason GM and Mazur JE (1996) Precipitating electrons: Dorrian GD, Breen AR, Davies JA, Rouillard AP, Whittaker IC, Davis CJ, Brown DS, evidence for effects on mesospheric odd nitrogen. Geophys Res Lett 23. Fallows RA, Harrison RA, Grande M (2010) Transient structures and stream doi:10.1029/96GL01787. interaction regions in the solar wind: results from EISCAT interplanetary Campbell BA (2002) Radar remote sensing of planetary surfaces. Cambridge scintillation, STEREO HI and Venus Express ASPERA-4 observations. Sol Phys University Press, ISBN 0-521-58308, QB603.S95 C36 265(1–2):207–231. doi:10.1007/s11207-010-9599-z Campbell BA, Campbell DB, Margot JL, Ghent RR, Nolan M, Chandler J, Carter LM, Dunkerton TJ, Delisi DP (1986) Evolution of potential vorticity in the winter Stacy NJS (2007) Focused 70-cm wavelength radar mapping of the moon. stratosphere of January-February, 1979. Geophys Res 91:1199–1208, diffusion IEEE Trans Geosci Remote Sens 45:4032–4042 region in the Earth’s magnetotail: The 2001–2005 Cluster observations Cane HV, McGuire RE, von Rosenvinge TT (1986) Two classes of solar energetic Dyer CS, Lei F, Clucas SN, Smart DF, Shea MA (2003) Calculations and particle events associated with impulsive and long-duration soft X-ray flares. observations of solar particle enhancements to the radiation environment at Astrophys J, Part 1 (ISSN 0004-637X) 301:448–459 aircraft altitudes. Adv Space Res 32:81–93. doi:10.1016/S0273-1177(03)90374-7 Carlson CW, And PRF, Watzin JG (1998) The Fast Auroral SnapshoT (FAST) Eastwood JP, Phan TD, Øieroset M, Shea MA (2010) Average properties of the Mission. Geophys Res Lett 25:12. doi:10.1029/98GL01592 magnetic reconnection ion and comparison with simulations. J Geophys Res Carslaw KJ, Harrison RG, Kirkby J (2002) Cosmic rays, clouds, and climate. Science 115:A08215. doi:10.1029/2009JA014962 298(5599):1732–1737. doi:10.1126/science.1076964 Ebel A (1984) Contributions of gravity waves to the momentum, heat and Cerruti AP, Kintner PM Jr, Gary DE, Mannucci AJ, Meyer RF, Doherty P, Coster AJ turbulent energy budget of the upper mesosphere and lower thermosphere. (2008) Effect of intense December 2006 solar radio bursts on GPS receivers. J Atmos Terr Phys 46(9):727–737 Space Weather 6:S10D07. doi:10.1029/2007SW000375 Echer E, Gonzalez WD, Guarnieri FL, Dal Lago A, Vieira LEA (2005) Introduction to Chau JL, Galindo F (2008) First definitive observations of meteor shower particles space weather. Adv Space Res 35(5):855–865 using a high-power large-aperture radar. Icarus 194(1):23–29. doi:10.1016/ Ecklund WL, Balsley BB (1981) Long-term observations of the arctic mesosphere j.icarus.2007.09.021 with the MST radar at Poker Flat, Alaska. J Geophys Res 86:7775–7780 Chelton DB, Schlax MG (1996) Global observations of oceanic Rossby waves. Emmert JT, Picone JM, Meier RR (2008) Thermospheric global average density Science 272:234–238. doi:10.1126/science.272.5259.234 trends, 1967–2007, derived from orbits of 5000 near-Earth objects. Geophys Chen CX, Wolf RA (1993) Interpretation of high-speed flows in the plasma sheet. Res Lett 35:L05101. doi:10.1029/2007GL032809 J Geophys Res 98:21409–21419 Emmert JT, Lean JL, Picone JM (2010) Record-low thermospheric density during Chen P, Holton JR, O’Neill A, Swinbank R (1994) Quasi-horizontal transport and the 2008 solar minimum. Geophys Res Lett 37:L12102. doi:10.1029/ mixing in the Antarctic stratosphere. J Geophys Res 99(D8):16851–16866 2010GL043671 McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 57 of 63

Enell CF, Kero A, Turunen E, Ulich T, Verronen PT, Seppala A, Marple S, Honary F, Graf H, Kirchner I, Perlwitz J (1998) Changing lower stratospheric circulation: the Senior A (2005) Effects of D-region heating studied with the Sodankyla Ion role of ozone and greenhouse gases. J Geophys Res 103:11,251–11,261 Chemistry Model. Ann Geophys 23(5):1575–1583 Gray LJ, Beer J, Geller M, Haigh JD, Lockwood M, Matthes K, Cubasch U, Erlykin AD, Laken BA, Sloan T, Wolfendale AW (2010) Cosmic rays and climate Fleitmann D, Harrison G, Hood L, Luterbacher J, Meehl GA, Shindell D, van change. 10th EMS Annual Meeting, 10th European Conference on Geel B, White W (2010) Solar influences on climate. Rev Geophys 48:RG4001. Applications of Meteorology (ECAM) Abstracts, Zürich, Switzerland, doi:10.1029/2009RG000282 id.EMS2010-790 Grydeland T, La Hoz C, Hagfors T, Blixt EM, Saito S, Strømme A, Brekke A (2003) Farley DT (1996) Incoherent scatter probing. In: Kohl H, Ruster R, Schlegel K Interferometric observations of filamentary structures associated with plasma (eds) Modern ionospheric science, 14th edn. European Geophysical instability in the auroral ionosphere. Geophys Res Lett 30(6):1338. Society, Katlenburg-Lindau, pp 415–439 doi:10.1029/2002GL016362 Fedrizzi M, Fuller-Rowell TJ, Codrescu MV (2012) Global Joule heating index derived Grydeland T, Blixt E, Løvhaug UP, Hagfors T, La Hoz C, Trondsen T (2004) from thermospheric density physics-based modelling and observations. Space Interferometric radar observations of filamented structures due to plasma Weather 10:S03001. doi:10.1029/2011SW00724 instabilities and their relation to dynamic auroral rays. Ann Geophys Feynman J, Gabriel SB (2000) On space weather consequences and predictions. 22:1115–1132 J Geophys Res 105(A5):10543–10564 Grydeland T, Lind FD, Erickson PJ, Holt JM (2005a) Software radar signal Fisher G, Kunches J (2011) Building resilience of the Global Positioning System to processing. Ann Geophys 23:109–121. doi:10.5194/angeo-23-109-2005 space weather. Space Weather 9:S12004. doi:10.1029/2011SW000718 Grydeland T, Chau JL, La Hoz C, Brekke A (2005b) An imaging interferometry Fisher M, O’Neill A, Sutton R (1993) Rapid descent of mesospheric air in the capability for the EISCAT Svalbard Radar. Ann Geophys 23:221–230 stratospheric polar vortex. Geophys Res Lett 20:1267–1270 Gustavsson B, Leyser TB, Kosch MJ, Rietveld MT, Steen A, Brandstrom BUE, Aso T Forbes JM (1982a) Atmospheric tides. J Geophys Res 87(A7):5222–5240. (2006) Electron gyroharmonic effects in ionization and electron acceleration doi:10.1029/JA087iA07p05241 during HF pumping in the ionosphere. Phys Rev Lett 97(195002). Forbes JM (1982b) Atmospheric tides: 2. The solar and lunar semidiurnal doi:10.1103/PhysRevLett.97.195002 components. J Geophys Res 87(A7):5241–5252. doi:10.1029/JA087iA07p05241 Gustavsson B, Kosch MJ, Senior A, Kavanagh AJ, Brandstrom U, Blixt EM (2008) Forme FRE (1993) A new interpretation on the origin of enhanced Combined EISCAT radar and optical multi-spectral and tomographic observations ionacoustic fluctuations in the upper ionosphere. Geophys Res Lett of black aurora. J Geophys Res 113:A06308. doi:10.1029/2007JA012999 20:2347–2350. doi:10.1029/93GL02490 Hagan ME, Forbes JM, Richmond A (2003) Atmospheric tides encyclopedia of Forte B, Smith ND, Mitchell CN, Da Dalt F, Panicciari T, Chartier AT, Stevanovic D, atmospheric sciences Vuckovic M, Kinrade J, Tong JR, Haggstrom I, Turunen E (2013) Comparison of Häggström I, Opgenoorth H, Williams PJS, Jones GOL, Schlegel KB (1989) temporal fluctuations in the total electron content estimates from EISCAT and Application of alternating codes for EISCAT observations during the GPS along the same line of sight. Ann Geophys 31:745–753 ERRRIS campaign for E-region plasma irregularities. J Atmos Terr Phys Foster JC, Holt JM, Musgrove RC, Evans DS (1986) Ionospheric convection 52(6–8):431–438 associated with discrete levels of particle precipitation. Geophys Res Lett Haigh JD (1996) The impact of solar variability on climate. Science 272:981–984 13:656 Haigh JD (1999) A GCM study of climate change in response to the 11-year solar Foster JC, et al. (2005a) Multiradar observations of the polar tongue of ionization. cycle. Quart J Roy Meteor Soc 125:871–892 J Geophys Res 110:A09S31. doi:10.1029/2004JA010928. Haigh JD (2003) The effects of solar variability on the earth’s climate. Philos Trans Foster JC, Coster AJ, Erickson PJ, Rideout W, Rich FJ, Immel TJ and Sandel BR Roy Soc London 361:95–111 (2005b) Redistribution of the stormtime ionosphere and the formation of a Hamilton K (1996) Comprehensive meteorological modelling of the middle plasmaspheric bulge. Inner magnetosphere interactions: new perspectives atmosphere: a tutorial review. J Atmos Sol Terr Phys 58:1591–1628 from imaging, Washington DC American Geophysical Union Geophysical Hamilton K (1999) Dynamical coupling of the lower and middle Monograph Series, v.159, Burch JL, Schulz M, Spence H, eds. American atmosphere: historical background to current research. J Atmos Sol Geophysical Union, 277–289, 2005. Terr Phys 61:73–84 Fritts DC, Luo Z (1995) Dynamical and radiative forcing of the summer Hanbaba R (1995) Statistical use of ionosonde data for IRI. Adv Space Res 15:17–22 mesopause circulation and thermal structure—1. Mean solstice conditions. Hansen RC (2009) Phased array antennas. John Wiley and Sons, Technology and J Geophys Res 100(D2):3119–3128 Engineering Series Fritts DC, van Zandt TE (1993) Spectral estimates of gravity wave and Hapgood MA (2011) Towards a scientific understanding of the risk from extreme momentum fluxes: part 1, energy dissipation, acceleration and constraints. space weather. Adv Space Res 47(12):2059–2072. doi:10.1016/ J Atmos Sci 50:3685–3694 j.asr.2010.02.007, ISSN 0273–1177 Fritts DC, Wang L, Werne J, Lund T, and Wan K (2009a) Gravity wave instability Hapgood MA, Harrison RA (1994) On the potential of interplanetary scintillation dynamics at high Reynolds numbers. Part I: Wave fieldevolution at large for predicting geomagnetic activity. Geophys Res Lett 21:637–640 amplitudes and high frequencies. J Atmos Sci 66:1126-1148. doi:10.1175/ Hartmann DL (1981) Some aspects of the coupling between radiation, chemistry, 2008JAS2726.1 and dynamics in the stratosphere. J Geophys Res 86, C10:9631–9640. Fritts DC, Wang L, Werne J, Lund T, and Wan K (2009b) Gravity wave instability doi:10.1029/JC086iC10p09631 dynamics at high Reynolds numbers. Part II: Turbulence evolution, structure, Hartmann DL, Wallace JM, Limpasuvan V, Thompson DWJ, Holton JR (2000) Can and anisotropy. J Atmos Sci 66:1149 –1171. doi:10.1175/2008JAS2727.1 ozone depletion and global warming interact to produce rapid climate Fuller-Rowell TJ, Rees D, Quegan S, Moffett RJ, Bailey GJ (1987) Interactions change? Proc Natl Acad Sci 92:1412–1417 between neutral thermospheric composition and the polar ionosphere using Havnes O, La Hoz C, Naesheim LI, Rietveld MT (2003) First observations of the PMSE a coupled ionosphere-thermosphere model. J Geophys Res 92:7744–7748 overshoot effect and its use for investigating the conditions in the summer Gabriel A, Peters D, Kirchner I, Graf HF (2007) Effect of zonally asymmetric ozone mesosphere. Geophys Res Lett 30(23):2229. doi:10.1029/2003GL018429 on stratospheric temperature and planetary wave propagation. Geophys Res Havnes O, LaHoz C, Biebricher A, Kassa M, Meseret T, Naesheim LI, Zivkovic T Lett 34:L06807. doi:10.1029/2006GL028998 (2004) Investigation of the mesospheric PMSE conditions by use of the new Gadsden M (1990) A secular change in noctilucent cloud occurrence. J Atmos overshoot effect. Phys Scripta T107:70 Terr Phys 52(4):247–251 Haynes PH, Marks CJ, McIntyre ME, Shepherd TE, Shine KP (1991) On the Gadsden M (1998) The north-west Europe data on noctilucent clouds: a survey”, “downward control” of extratropical adiabatic circulation by eddy-induced J Atmos Sol Terr Phys 60:163-1174. doi:10.1016/S1364-6826(98)00072-8 mean zonal forces. J Atmos Sci 48:651–678 Garcia G, Forme F (2009) A kinetic model of the ionospheric return currents. Hedin AE (1987) The MSIS-86 thermospheric model. J Geophys Res 92:4649 Geophys Res Lett 36:L20106. doi:10.1029/2009GL039972 Heinselman CJ, Nicolls MJ (2008) Bayesian approach to electric field and E-region Goncharenko L, Zhang SR (2008) Ionospheric signatures of sudden stratospheric neutral wind estimation with the Poker Flat Advanced Modular Incoherent warming: ion temperature at middle latitude. Geophys Res Lett 35:L21103. Scatter Radar. Radio Sci 43:RS5013. doi:10.1029/ 2007RS003805 doi:10.1029/2008GL035684 Henderson MG, Reeves GD, Murphree JS (1998) Are north-south structures Gordon W (1958) Incoherent scattering of radio waves by free electrons with an ionospheric manifestation of bursty bulk flows? Geophys Res Lett applications to space exploration by radar. Proc IRE 46(11):1824–1829 25:3737–3740 McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 58 of 63

Hervig ME, Gordley LL, Russell JM III, Bailey SM (2009) SOFIE PMC observations Kero A, Bosinger T, Pollari P, Turunen E, Rietveld MT (2000) First EISCAT during the northern summer of 2007. J Atmos Sol Terr Phys 71:331–339. measurement of electron-gas temperature in the artificially heated D-region doi:10.1016/j.jastp.2008.08.010 ionosphere. Ann Geophys 18:1210–1215 Hewish A (1989) A user’s guide to scintillation. J Atmos Terr Phys 51:743–750 Kero J, Szasz C, Pellinen-Wannberg A, Wannberg G, Westman A, Meisel DD Hill RJ, Gibson-Wilde DE, Werne JA, Fritts DC (1999) Turbulence-induced fluctuations (2008a) Three-dimensional radar observation of a submillimeter meteoroid in ionization and application to PMSE. Earth Planets Space 51:499–513 fragmentation. Geophys Res Lett 35:L04101. doi:10.1029/2007GL032733 Hocke K, Schlegel K (1996) A review of atmospheric gravity waves and travelling Kero J, Szasz C, Wannberg G, Pellinen-Wannberg A, Westman A (2008b) On the ionospheric disturbances: 1982–1995. Ann Geophys 14:917–940 meteoric head echo radar cross section angular dependence. Geophys Res Hocking WK (1996) Dynamical coupling processes between the middle Lett 35:L07101. doi:10.1029/2008GL033402 atmosphere and lower ionosphere. J Atmos Sol Terr Phys 58:735–752 Kero J, Szasz C, Nakamura T, Meisel DD, Ueda M, Fujiwara Y, Terasawa T, Holt JM, Zhang S-R (2008) Long-term temperature trends in the ionosphere Miyamoto H, Nishimura K (2011) First results from the 2009–2010 MU radar above Millstone Hill. Geophys Res Lett 35:L05813. doi:10.1029/2007GL031148 head echo observation programme for sporadic and shower meteors: the Holton JR (1982) The role of gravity wave induced drag and diffusion in the Orionids 2009. Mon Not R Astron Soc 416:2550–2559. doi:10.1111/j.1365- momentum budget of the mesosphere. J Atmos Sci 39:791–799 2966.2011.19146.x Holton JR, Haynes PH, Mcintyre ME, Douglass AR, Rood RB, Pfister L (1995) Kersley L, Pryse SE, Wheadon N (1988) Small scale irregularities associated with a Stratosphere-toposphere exchange. Rev Geophys 33:403–439 high latitude electron density gradient: scintillation and EISCAT observations. Hooke WH (1986) Gravity waves. In: Ray PS (ed) Mesoscale meteorology and J Atmos Terr Phys 50(6):557–563 forecasting. Amer Meteor Soc. 272–288. Kessler DJ, Cour-Palais BG (1978) Collision frequency of artificial satellites: the Hunsucker RD (1982) Atmospheric gravity waves generated in the high-latitude creation of a debris belt. J Geophys Res 83:2637–2646 ionosphere: a review. Rev Geophys 20(2):293–315. doi:10.1029/RG020i002p00293 Kinsey R (1997) Phased array beam spoiling technique. Antennas and Hunsucker RD and Hargreaves JK (2002) The high-latitude ionosphere and its Propagation Society International Symposium, IEEE. 1997 Digest effects on radio propagation. Cambridge University Press (Volume:2), 698 – 701 Huuskonen A, Lehtinen MS (1996) General incoherent scatter analysis and Kintner PM, Humphreys T, Hinks J (2009a) GNSS and ionospheric GUISDAP error estimates valid for high signal strengths. J Atmos Terr scintillation—how to survive the next solar maximum, inside GNSS. July/ Phys 58:435–464 August 2009, http://www.insidegnss.com/node/1579. Inan US, Golkowski M, Carpenter DL, Reddell N, Moore RC, Bell TF, Paschal E, Kintner PM, O’Hanlon B, Gary DE, Kintner PMS (2009b) Global positioning system Kossey P, Kennedy E, Meth SZ (2004) Multi-hop Whistler-mode ELF/VLF and solar radio burst forensics. Radio Science 44:RS0A08. doi:10.1029/ signals and triggered emissions excited by the HAARP HF Heater. Geophys 2008RS004039 Res Let 31:L24805. doi:10.1029/2004GL021647 Kirkwood S, Nilsson H (2000) High-latitude sporadic-E and other thin layers—the Jackman CH, Fleming EL, Vitt FM (2000) Influence of extremely large solar proton role of magnetospheric electric fields. Space Sci Rev 91.3-4(2000):579–613 events in a changing stratosphere. J Geophys Res 105:11,659–11,670 Kirkwood S, Barabash V, Belova E, Nilsson H, Rao TN, Stebel K, Osepian A, Chilson Jackman CH et al (2005a) Neutral atmospheric influences of the solar proton PB (2002) Polar mesosphere winter echoes during solar proton events. Adv events in October-November 2003. J Geophys Res 110:A09S27. Polar Upper Atmos Res 16:111–125 doi:10.1029/2004JA010888 Kirkwood S, Belova E, Blum U, Croskey C, Dalin P, Fricke KH, Goldberg RA, Jackman CH et al. (2005b) The influence of the several very large solar proton Mitchell JD, Schmidlin F (2006a) Polar mesosphere winter echoes during events in years 2000–2003 on the neutral middle atmosphere. Adv Space MaCWAVE. Ann Geophys 24:1245–1255 Res 35(3). doi:10.1016/j.asr.2004.09.006. Kirkwood S, Chilson P, Belova E, Dalin P, Häggström I, Rietveld MT, Singer W Jakowski N (2005) Ionospheric GPS radio occultation measurements on board (2006b) Infrasound—the cause of strong polar mesosphere winter echoes? CHAMP. GPS Solutions 9:88–95. doi:10.1007/s10291-005-0137-7 Ann Geophys 24:475–491 Janhunen P (1996) GUMICS-3—a global ionosphere-magnetosphere coupling Kirkwood S, Dalin P, Rechou A (2008) Noctilucent clouds observed from the simulation with high ionospheric resolution. ESA Symposium Proceedings on UK and Denmark—trends and variations over 43 years. Ann Geophys ‘Environmental Modelling for Space-based applications’. ESTEC, Noordwijk, 26:1243–1254 NL, 18–20 Sep 1996, ESA SP-392 Kirkwood S, Belova E, Satheesan K, Narayana Rao T, Rajendra Prasad T, Janhunen P, Palmroth M, Laitinen T, Honkonen I, Juusola L, Facsko G, Pulkkinen Satheesh Kumar S (2010a) Fresnel scatter revisited—comparison of 50 TI (2012) The GUMICS-4 global MHD magnetosphere–ionosphere coupling MHz radar and radiosondes in the Arctic, the Tropics and Antarctica. simulation. J Atmos Sol-Terr Phys 80:48–59. doi:10.1016/j.jastp.2012.03.006 Ann Geophys 28:1993–2005. doi:10.5194/angeo-28-1993-2010 Johansson J, Borg J, Lindgren T, Beckman C, Isaaksson L, van Zwol K, Berglund P, Kirkwood S, Mihalikova M, Rao TN, Satheesan K (2010b) Turbulence associated with Strandell J, Knuutinen K, Ek M (2014) EISCAT_3D Preparatory Phase Project: mountain waves over northern Scandinavia—a case study using the ESRAD Deliverable 8.5: Report on all WP8 activities., https://www.eiscat3d.se/sites/ VHF radar and the WRF mesoscale model. Atmos Chem Phys 10:3583–3599 default/files/WP8_Final_report.pdf Klinkrad, H (2006) Space debris models and risk analysis. Springer Kaasalainen M, Lamberg L (2006) Inverse problems of generalized projection Klinkrad H (2010) Space debris. In: Encylopedia of aerospace engineering. John operators. Inverse Problems 22:749 Wiley and Sons, doi:10.1002/97804706886652.eae325 Kato S (1989) Atmospheric tides—a review. J Atmos Terr Phys 43(5–6):491–493 Klinkrad H, Jehn R (1992) The space-debris environment of the earth. ESA J 16(1):1–11 Kato S (2005) Middle atmosphere research and radar observation. Proc Jpn Acad Knipp D, Eriksson S, Kilcommons L, Crowley G, Lei J, Hairston M, Drake K (2011) Series B 81(8):306–320 Extreme Poynting flux in the dayside thermosphere: examples and statistics. Kauristie K, Sergeev VA, Amm O, Kubyshkina M, Jussila J, Liou K (2003) Bursty bulk Geophys Res Lett 38:L16102. doi:10.1029/2011GL048302 flow intrusion to the inner plasma sheet as inferred from auroral Knudsen DJ, Donovan EF, Cogger LL, Jackel BJ, Shaw WD (2001) Width and observations. J Geophys Res 108, A1:1040–1048 structure of mesoscale optical auroral arcs. Geophys Res Lett 28:705–708 Kavanagh AJ, Honary F, Rietveld MT, Senior A (2006) First observations of the Kosch MJ, Rietveld MT, Kavanagh AJ, Davis CJ, Yeoman TK, Honary F, Hagfors T artificial modulation of polar mesospheric winter echoes. Geophys Res Lett (2002) High-latitude pump-induced optical emissions for frequencies close to 33:L19801. doi:10.1029/2006GL027575 the third electron gyro-harmonic. Geophys Res Lett 29(23):2112–2115 Keating GM, Tolson RH, Bradford MS (2000) Evidence of long term global decline Kosch MJ, Rieveld MT, Senior A, McCrea IW, Kavanagh AJ, Isham B, Honary F in the Earth’s thermospheric densities apparently related to anthropogenic (2004) Novel artificial optical annular structures in the high latitude effects. Geophys Res Lett 27:1523–1526 ionosphere over EISCAT. Geophys Res Lett 31:L12805. doi:10.1029/ Keiling A (2009) Alfvén waves and their roles in the dynamics of the earth’s 2004GL019713 magnetotail: a review. Space Sci Rev 142:73–156. doi:10.1007/s11214-008-9463-8 Kosch MJ, Pedersen T, Mishin E, Starks M, Gerken-Kendall E, Sentman D, Oyama S, Kelley MC (2009) The Earth’s ionosphere: plasma physics and electrodynamics. Watkins B (2007) Temporal evolution of pump beam self-focusing at HAARP. Academic Press J Geophys Res 112:A08304. doi:10.1029/2007JA012264 Kelley MC, Arce TL, Salowey J, Sulzer M, Armstrong WT, Carter M, Duncan L Kosch MJ, Ogawa Y, Rietveld MT, Nozawa S, Fujii R (2010) An analysis of (1995) Density depletions at the 10-m scale induced by the Arecibo heater. pump-induced artificial ionospheric ion upwelling at EISCAT. J Geophys Res J Geophys Res 100:17367–17376 115:A12317. doi:10.1029/2010JA015854 McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 59 of 63

Kosch MJ, Yiu I, Anderson C, Tsuda T, Ogawa Y, Nozawa S, Aruliah AL, Howells V, Limpasuvan V, Hartmann DL (2000) Wave-maintained annular modes of climate Baddeley L, McCrea IW, Wild JA (2011a) Meso-scale Joule heating near an variability. J Clim 13:4414–4429. doi:10.1175/1520-0442 auroral arc and ion-neutral collision frequency in the polar cap E-region. Limpasuvan V, Thompson DWJ, Hartmann DL (2004) The life cycle of the J Geophys Res 116:A05321. doi:10.1029/2010JA016015 Northern Hemisphere sudden stratospheric warmings. J Clim 17:2584–2596 Kosch MJ, Mjolhus E, Ashrafi M, Rietveld MT, Yeoman TK, Nozawa S (2011b) Lind FD (2011) Next generation space science with the geospace array., Angular dependence of pump-induced topside plasma turbulence at EISCAT. http://www.hao.ucar.edu/CommunityWorkshopGroundBasedSolarResearch/ J Geophys Res 116:A03322. doi:10.129/2010JA016014 whitePages/Next%20Generation%20Space%20Science%20with%20the Kulmala M, Riipinen I, Nieminen T, Hulkkonen M, Sogacheva L, Manninen HE, %20Geospace%20Array.pdf Paasonen P, Petaja T, Dal Maso M, Aalto PP, Viljanen A, Usoskin I, Vainio R, Lind FD, Grydeland T, Erickson PJ, Holt JM and Sahr JD (2002) Advanced techniques Mirme S, Mirme A, Minikin A, Petzold A, Horrak U, Plass-Dulmer C, Birmili W, for incoherent scatter radar. Proceedings of the 2002 URSI General Assembly, Kerminen VM (2009) Atmospheric data over a solar cycle: no connection Maastricht, NL, http://www.ursi.org/Proceedings/ProcGA02/papers/p0943.pdf between galactic cosmic rays and new particle formation. Atmos Chem Phys Lindzen RS (1979) Atmospheric tides. Annu Rev Earth Planet Sci 7:199–225. Discuss 9:21525–21560 doi:10.1146/annurev.ea.07.050179.001215 Kurihara J, Ogawa Y, Oyama S, Nozawa S, Tsutsumi M, Hall CM, Tomikawa Y, Fujii Lindzen RS (1981) Turbulence and stress owing to gravity wave and tidal R (2010) Links between a stratospheric sudden warming and thermal breakdown. J Geophys Res 86(CI0):9707–9714 structures and dynamics in the high-latitude mesosphere, lower Lindzen RS (1984) Gravity waves in the mesosphere. In: Holton JR, Matsuno T thermosphere, and ionosphere. Geophys Res Lett 37:L13806. doi:10.1029/ (eds) Dynamics of the middle atmosphere., Terrapub, 3–18 2010GL043643 Lindzen RS (2007) Taking greenhouse warming seriously. Energy Environ 18:937–950 Kuroda Y, Kodera K (1999) Role of planetary waves in the stratosphere– Lübken FJ (1999) Thermal structure of the Arctic summer mesosphere. J Geophys troposphere coupled variability in the northern Hemisphere winter. Geophys Res 104:9135–9149 Res Lett 26:2375–2378 LübkenFJ,StrelnikovB,RappM,SingerW,LatteckR,BrattliA,HoppeUP, La Hoz C and Belyey V (2011) EISCAT aperture synthesis imaging for the Friedrich M (2006) The thermal and dynamical state of the atmosphere EISCAT_3D project. URSI General Assembly and Science Symposium, during polar mesosphere winter echoes. Atmos Chem Phys 6:13–24 doi:10.1109/URSIGASS.2011.6050894, ISBN 978-1-4244-5117-3 Lübken FJ, Singer W, Latteck R, Strelnikova I (2007) Radar measurements of La Hoz C and Havnes O (2008) Artificial modification of polar mesospheric winter turbulence, electron densities, and absolute reflectivities during polar echoes with an RF heater: do charged dust particles play an active role? mesosphere winter echoes (PMWE). Adv Space Res 40(6):758–764 J Geophys Res. 113: doi:10.1029/2008JD010460 Lübken FJ, Lautenbach J, Hoffner J, Rapp M, Zecha M (2009) First continuous Labitzke K (1972) Temperature changes in the mesosphere and stratosphere temperature measurements within polar mesospheric summer echoes. connected with circulation changes in winter. J Atmos Sci 29:756–766 J Atmos Terr Phys 71:453–463 Labitzke K (1981) Stratospheric-mesospheric midwinter disturbances: a summary Lühr H, Rother M, Köhler W, Ritter P, Grunwaldt L (2004) Thermospheric of observed characteristics. J Geophys Res 86:9665–9678 upwelling in the cusp region: evidence from CHAMP observations. Geophys Labitzke K (1987) Sunspots, the QBO, and the stratospheric temperature in the Res Lett 31:L06805. doi:10.1029/2003GL019314 north polar-region. Geophys Res Lett 14:535–537 Maggs JE, Davis TN (1968) Measurements of the thicknesses of auroral structures. Labitzke K (2001) The global signal of the 11-year sunspot cycle in the strato- Planet Space Sci 16:205–209 sphere: differences between solar maxima and minima. Meteor Z 10:83–90 Mann I (1995) Spatial distribution and orbital properties of interplanetary dust at Labitzke K (2003) The global signal of the 11-year sunspot cycle in the atmosphere: high latitudes. Space Sci Rev 72:477–480 when do we need the QBO? Meteor Z 12:209–216 Mann I, Häggström I, Tjulin A, Nylén M (2013) The e-infrastructure for Labitzke K (2004) On the signal of the 11-year sunspot cycle in the stratosphere and its EISCAT_3D, EISCAT_3D Preparatory Phase Report. https://www.eiscat3d.se/ modulation by the quasi-biennial oscillation. J Atmos Sol Terr Phys 66:1151–1157 content/e-infrastructure-eiscat3d Labitzke K (2005) On the solar cycle-QBO relationship: a summary. J Atmos Sol Markkanen J, Lehtinen MS, Huuskonen A and Vaananen A (2002) Measurements of Terr Phys 67:45–54 small-size debris with backscatter of radio waves. Final Report, ESOC Contract Lastovicka J, Solomon SC, Qian L (2012) Trends in the neutral and ionized upper No. 13945/99/D/CD atmosphere. Space Sci Rev 168:113–145. doi:10.1007/s11214-011-9799-3 Markkanen J, Lehtinen MS, Landgraf M (2005) Real-time space debris monitoring Latteck R, Singer W, Rapp M, Renkwitz T (2010) MAARSY—the new MST radar on with EISCAT. Adv Space Res 35:1197–1208 Andøya, Norway. Adv Radio Sci 8:219–224. doi:10.5194/ars-8-219-2010 Markkanen J, Jehn R, Krag H (2009) EISCAT space debris during the IPY—a 5000 Lehtinen MS, Huuskonen A (1996) General incoherent scatter analysis and hour campaign, Proc. 5th ESA space debris conference GUISDAP. J Atmos Terr Phys 58:435–452 Marklund G (1984) Auroral arc classification scheme based on the observed Lehtinen MS, Huuskonen A, Markkanen M (1998) Randomization of alternating arc-associated electric field patterns. Planet Space Sci 32:193 codes: improving incoherent scatter measurements by reducing correlations Marsh N, Svensmark H (2000a) Cosmic rays, clouds, and climate. Space Sci Rev of gated autocorrelation function estimates. Radio Sci 32/6:2271–2282 94(1–2):215–230 Lehtinen MS, Virtanen II, Vierinen J (2008) Fast comparison of IS radar code Marsh N, Svensmark H (2000b) Low cloud properties influenced by cosmic rays. sequences for lag profile inversion. Ann Geophys 26:2291–2301 Phys Rev Lett 85(23):5004–5007 Lehtinen MS, Damtie B, Piiroinen P, Orispää M (2009) Perfect and almost perfect McIntyre ME (1982) How well do we understand the dynamics of stratospheric pulse compression codes for range spread radar targets. Inverse Prob warmings? J Meteor Soc Japan 60:37–65 Imaging 3(3):465–486 McIntyre ME (1989) On dynamics and transport near the polar mesopause in Lehtinen MS, Virtanen II and Orispää M (2014) EISCAT_3D measurement methods summer. J Geophys Res 94(D12):14,617–14,628 handbook, Deliverable 6.7 of the EISCAT_3D Preparatory Phase Project. McLandress C (1998) On the importance of gravity waves in the middle atmosphere https://www.eiscat3d.se/sites/default/files/Handbook_D67v099.pdf and their parameterization in general circulation models. J Atmos Sol Terr Phys Lei F, Hands A, Clucas S, Dyer C, Truscott P (2006) Improvement to and 60:1357–1384 validations of the QinetiQ Atmospheric Radiation Model (QARM). IEEE Trans McNamara LF, Decker DT, Welsh JA, Cole DG (2007) Validation of the Utah Nucl Sci 53(4):1851–1858 State University Global Assimilation of Ionospheric Measurements (GAIM) Leitinger R (1996) Tomography. In: Kohl H, Ruster R, Schlegel K (eds) Modern ionospheric model predictions of the maximum usable frequency for a 3000 km science. European Geophysical Society, Katlenburg-Lindau, pp 346–370 circuit. Radio Sci 42(RS3015):17. doi:10.1029/2006RS003589 Lewis HG, Saunders A, Swinerd GG, Newland RJ (2011) Effect of thermospheric McPherron RL, Weygand JM, Hsu TS (2008) Response of the Earth’s contraction on remediation of the near-Earth space debris environment. magnetosphere to changes in the solar wind. J Atmos Sol Terr Phys J Geophys Res 116:A2. doi:10.1029/2011JA016482 70:303–315 Leyser TB, Wong AY (2009) Powerful electromagnetic waves for active environmental Mendonca JT, Thidé B, Then H (2009a) Stimulated Raman and Brillouin research in geospace. Rev Geophys 47. doi:10.1029/2007RG000235, 2009a. backscattering of collimated beams carrying orbital angular momentum. Leyser T, Norin L, McCarrick M, Pedersen TR, Gustavsson B (2009) Radio pumping Phys Rev Lett 102:185005. doi:10.1103/PhysRevLett.102.185005 of ionospheric plasma with orbital angular momentum. Phys Rev Lett Mendonca JT, Ali S, Thidé B (2009b) Plasmons with orbital angular momentum. 102:065004. doi:10.1103/PhysRevLett 102.065004 Phys Plasmas 16:112103. doi:10.1063/1.3261802 McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 60 of 63

Meriwether JW, Gardner CS (2000) A review of the mesosphere inversion layer Ostgaard N, Germany G, Stadsnes J, Vondrak RR (2002) Energy analysis of phenomenon. J Geophys Res 105:12405–12416 substorms based on remote sensing techniques, solar wind measurements, Mertens CJ, Kress BT, Wiltberger M, Blattnig SR, Slaba TS, Solomon SC, Engel M and geomagnetic indices. JGR 107:1233. doi:10.1029/2001JA002002 (2010) Geomagnetic influence on aircraft radiation exposure during a solar Ostro SJ (1993) Planetary radar astronomy. Rev Mod Phys 65:1235–1279 energetic particle event in October 2003. Space Weather 8:S03006. Pancheva D et al (2008) Planetary waves in coupling the stratosphere and doi:10.1029/2009SW000487 mesosphere during the major stratospheric warming in 2003/2004. J Michell RG, Lynch KA, Heinselman CJ, Stenbaek‐Nielsen HC (2008) PFISR Geophys Res 113:D12105. doi:10.1029/2007JD009011 nightside observations of naturally enhanced ion acoustic lines, and their Partamies N, Syrjäsuo M, Donovan E, Connors M, Charrois D, Knudsen D, relation to boundary auroral features. Ann Geophys 26:3623–3639. Kryzanowsky Z (2010) Observations of the auroral width spectrum at doi:10.5194/angeo-26-3623-2008 kilometre-scale size. Ann Geophys 28:711–718 MiddletonHR,PryseSE,WoodAG,BalthazorRL(2008)Theroleofthe Pellinen-Wannberg A, Wannberg G (1994) Meteor observations with the EISCAT tongue-of-ionisation in the formation of the poleward wall of the main UHF incoherent scatter radar. J Geophys Res 99:11379–11390 trough in the European post-midnight sector. J Geophys Res Pellinen-Wannberg A, Murad E, Gustavsson B, Brändström U, Enell CF, Roth C, 113(A2):A02306. doi:10.1029/2007JA012631 Williams IP, Steen Å (2004) Optical observations of water in Leonid meteor Mikhailov AV, Depuev VH, Depueva AH (2007) Short-term foF2 forecasting: trails. Geophys Res Lett 31:L03812. doi:10.1029/2003GL018785 present-day state of art. In: J. Lilensten (ed) Space weather—research Perlwitz J, Harnik N (2003) Observational evidence of a stratospheric influence on the towards applications in Europe. Springer Astrophysics and Space Science troposphere by planetary wave reflection. J Clim 16:3011–3026 Library, 344. ISBN-10 1-4020-5445-9 Perlwitz J, Harnik N (2004) Downward coupling between the stratosphere and Mitchell CN, Alfonsi L, De Franceschi G, Lester M, Romano V, Wernik AW (2005) GPS troposphere: the relative roles of wave and zonal mean processes. J Clim TEC and scintillation measurements from the polar ionosphere during the 17:4902–4909. doi:10.1175/JCLI-3247.1 October 2003 storm. Geophys Res Lett 32:L12S03. doi:10.1029/2004GL021644 Pertsev N, Dalin P, Perminov V, Romejko V, Dubietis A, Balcunas R, Cernis K, Mozer FS, Cattell CA, Hudson MK, Lysak RL, Temerin M, Torbert RB (1980) Zalcik M (2014) Noctilucent clouds observed from the ground: sensitivity Satellite measurements and theories of low altitude auroral particle to mesospheric parameters and long-term time series. Earth Planets acceleration. Space Sci Rev 27:155–211 Space 66:98. doi:10.1186/1880-5981-66-98 Murad E, Williams IP (2002) Meteors in the Earth’s atmosphere. Cambridge PettengillGH,EliasonE,FordPG,LoriotGB,MasurskyH,McGillGE(1980) University Press, Cambridge, UK Pioneer Venus radar results altimetry and surface properties. J Geophys Nakamura T, Tsuda T, Yamamoto M, Fukao S, Kato S (1993) Characteristics of Res 85:8261–8270 gravity waves in the mesosphere observed with the middle and upper Peymirat C, Fontaine D (1994) Numerical simulation of magnetospheric atmosphere radar 1: momentum flux. J Geophys Res 98(D5):8899–8910. convection including the effect of field-aligned currents and electron doi:10.1029/92JD02978 precipitation. J Geophys Res 103:11155–11176 Nakamura R, Amm O, Laakso H, Draper NC, Lester M, Grocott A, Klecker B, Pfister L, Stanley S, Loewenstein M, Bowen S, Legg M (1993) Mesoscale McCrea IW, Balogh A, Rème H, André M (2005) Localized fast flow disturbances in the tropical stratosphere excited by disturbance observed in the plasma sheet and in the ionosphere. Ann convection—observations and effects on the stratospheric momentum Geophys 23:553–566. doi:10.5194/angeo-23-553-2005 budget. J Atmos Sci 50(8):1058–1075 Nilsson H, Sergienko TI, Ebihara Y, Yamauchi M (2005) Quiet-time mid-latitude Picardi G et al. (2005) Radar soundings of the subsurface of Mars. Science trough: influence of convection, field-aligned currents and proton precipitation. 310:1925–1928 Ann Geophys 23:3277–3288 Pierce JR, Adams PJ (2009) Can cosmic rays affect cloud condensation nuclei by Nygrén T, Jalonen L, Oksman J, Turunen T (1984) The role of electric field and neutral altering new particle formation rates. Geophys Res Lett 36:L09820. winddirectionintheformationofsporadic E layers. J Atmos Terr Phys 46:373–381 doi:10.1029/2009GL037946 Nygrén T, Aikio AT, Kuula R, Voiculescu M (2011) Electric fields and neutral winds Pirjola R, Kauristie K, Lappalainen H, Viljanen A, Pulkkinen A (2005) Space weather from monostatic incoherent scatter measurements by means of stochastic risk. Space Weather 3:S02A02. doi:10.1029/2004SW000112 inversion. J Geophys Res 116:A05305. doi:10.1029/2010JA016347 Pitkänen T, Aikio AT, Amm O, Kauristie K, Nilsson H, Kaila KU (2011) EISCAT-Cluster Odenwald S, Green J, Taylor W (2006) Forecasting the impact of an 1859-calibre observations of quiet-time near-Earth magnetotail fast flows and their signatures superstorm on satellite resources. Adv Space Res 38(2):280–297. doi:10.1016/ in the ionosphere. Ann Geophys 29:299–319 j.asr.2005.10.046, ISSN 0273–1177 Prikryl P, Jayachandran PT, Mushini SC, Chadwick R (2011) Climatology of GPS Ogawa Y, Häggström I, Buchert SC, Oksavik K, Nozawa S, Hirahara M, van Eyken phase scintillation and HF radar backscatter for the high-latitude ionosphere AP, Aso T, Fujii R (2009) On the source of the in the polar topside under solar minimum conditions. Ann Geophys 29:377–392. doi:10.5194/ ionosphere: first results from the EISCAT Svalbard Radar. Geophys Res Lett angeo-29-377 36:L24103. doi:10.1029/2009GL041501 Prusa JM, Smolarkiewicz PK, Garcia RR (1996) Propagation and breaking at high Ogawa Y, Buchert SC, Sakurai A, Nozawa S, Fujii R (2010) Solar activity altitudes of gravity waves excited by tropospheric forcing. J Atmos Terr Phys dependence of ion upflow in the polar ionosphere observed with the 53:2186–2216 European Incoherent Scatter (EISCAT) Tromsø UHF radar. J Geophys Res Pryse SE, Whittick EL, Aylward AD, Middleton HR, Brown DS, Lester M, Secan JA 115:A07310 (2009) Modelling the tongue‐of‐ionisation using CTIP with SuperDARN Ogawa Y, Buchert SC, Häggström I, Rietveld MT, Fujii R, Nozawa S, Miyaoka H electric potential input: verification by radiotomography. Ann Geophys (2011) On the statistical relation between ion upflow and naturally enhanced 27:1139–1152 ion‐acoustic lines observed with the EISCAT Svalbard Radar. J Geophys Res Pudovkin MI (2004) Influence of solar activity on the lower atmosphere state. Int 116:A03313. doi:10.1029/2010JA015827 J Geomagn Aeron 5:GI2007 Ogawa Y, Motoba T, Buchert SC, Häggström I, Nozawa S (2014) Upper Pulkkinen T (2007) Space weather: terrestrial perspective. Living Rev Solar Phys atmosphere cooling over the past 33 years. Geophys Res Lett 41:5629–5635. 4:1, http://www.livingreviews.org/lrsp-2007-1 doi:10.1002/2014GL060591 Pulkkinen A, Hesse M, Habib S, Van der Zel L, Damsky B, Policelli F, Fugate D, Øieroset M, Phan TD, Lin RP, Sonnerup B (2000) Walén and variance analyses of Jacobs W, Creamer E (2010) Solar shield: forecasting and mitigating space high-speed flows observed by Wind in the midtail plasma sheet: evidence weather effects on high-voltage power transmission systems. Nat Hazards for reconnection. J Geophys Res 105(A11):25, 247–25, 263 53:333–345. doi:10.1007/s11069-009-9432-x Oksavik K, Moen JI, Carlson HC (2004) High-resolution observations of the Qian L, Solomon SC, Roble RG, Kane TJ (2008) Model simulations of global small-scale flow pattern associated with a poleward moving auroral form change in the ionosphere. Geophys Res Lett 35:L07811. doi:10.1029/ in the cusp. Geophys Res Lett 31:L11807. doi:10.1029/2004GL019838 2007GL033156 Olivero JJ, Thomas GE (2001) Evidence for changes in greenhouse gases in Qian L, Laštovicka J, Solomon SC, Roble RG (2011) Progress in observations and the mesosphere. Adv Space Res 28(7):931–936 simulations of global change in the upper atmosphere. J Geophys Res Opgenoorth HJ, Lockwood M (1997) Opportunities for magnetospheric research 111:A00H03. doi:10.1029/2010JA016317 with coordinated cluster and ground-based observations. Space Sci Rev Randall CE et al. (1998) Polar Ozone and Aerosol Measurement (POAM) II 79:599–637 stratospheric NO2, 1993–1996. J Geophys Res 103:28,361–28,371 McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 61 of 63

Rapp M, Lübken FJ (2003) On the nature of PMSE: electron diffusion in the Ruohoniemi JM, Greenwald RA (2005) Dependencies of high-latitude plasma vicinity of charged particles revisited. J Geophys Res 108(D8):8437 convection: consideration of interplanetary magnetic field, seasonal, and Rapp M, Lübken FJ, Hoffmann P, Latteck R, Baumgarten G, Blix TA (2003a) PMSE universal time factors in statistical patterns. J Geophys Res 110, A09204. dependence on aerosol charge number density and aerosol size. J Geophys doi:10.1029/2004JA010815 Res 108(D8):8441. doi:10.1029/2002JD002650 Rusch DW, Gerard JC, Solomon S, Crutzen PJ, Reid GC (1981) The effect of Rapp M, Lübken FJ, Blix T (2003b) The role of charged ice particles in the creation of particle precipitation events on the neutral and ion chemistry of the middle PMSE: a review of recent developments. Adv Space Res 31(9):2033–2043 atmosphere—I. Odd nitrogen. Planet Space Sci 29:767–774 Rapp M, Strelnikova I, Latteck R, Hoffman P, Hoppe UP (2008) Polar mesosphere Rüster R, Nastrom GD, Schmidt G (1998) High-resolution VHF radar summer echoes (PMSE) studied at Bragg wavelengths of 2.8 m, 67 cm, and measurements in the troposphere with a vertically pointing beam. 16 cm. J Atmos Sol Terr Phys 70(7):947–961 J Appl Meteorol 37:1522–1529 Reames DV (1990) Energetic particles from impulsive solar flares. Astrophys J Rycroft MJ, Israelsson S, Price C (2000) The global atmospheric electric circuit, Suppl Ser 73:235–251, ISSN 0067–0049 solar activity and climate change. J Atmos Terr Phys 62:1563–1576 Rees MH (1989) Physics and chemistry of the upper atmosphere. Cambridge Rycroft MJ, Harrison RG, Nicoll KA, Mareev EA (2002) An overview of Earth’s University Press global electric circuit and atmospheric conductivity. Space Sci Rev 137:1–4. Reid GC (1989) Ion chemistry of the cold summer mesopause region. J Geophys doi:10.1007/s11214-008-9368-6, 83–105 Res 94:14653–14660. doi:10.1029/JD094iD12p14653 Sato K, Tsutsumi M, Sato T, Nakamura T, Saito A, Tomikawa Y, Nishimura K, Reid GC, Solomon S, Garcia RR (1991) Response of the middle atmosphere Yamagishi H and Yamanouchi T (2010) Current status of the Program of to the solar proton events of August–December, 1989. Geophys Res Lett the Antarctic Syowa MST/IS radar (PANSY): toward a new era of 18:1019–1022 Antarctic atmosphere research. Presented at the 74th EISCAT Council Reinisch BW, Huang X, Song P, Sales GS, Fung SF, Green JL, Gallagher DL, Vaslyunas meeting, Tachikawa, Tokyo, June 2010. VM (2001a) Plasma density distribution along the magnetic field: RPI Saunders RS, Pettengill GH, Arvidson RE, Sjogren WL, Johnson WTK, Pieri L (1990) observations from IMAGE. Geophys Res Lett 28. doi:10.1029/2001GL013684 The Magellan Venus Radar Mapping Mission. J Geophys Res 95:8339–8355 Reinisch BW, Haines DM, Sales GS, Benson RF, Green JL, Taylor WWL (2001b) Scherliess L, Thompson DC, Schunk RW (2009) Ionospheric dynamics and Radio sounding in space: magnetosphere and topside ionosphere. J Atmos drivers obtained from a physics-based data assimilation model. Radio Sci Sol Terr Phys 63:87–98 44:Artn Rs0a32. doi:10.1029/2008rs004068 Richmond AD (1978) Gravity wave generation, propagation, and dissipation in Schlatter NM, Grydeland T, Ivchenko N, Belyey V, Sullivan J, La Hoz C, Blixt M the thermosphere. J Geophys Res 83(A9):4131–4145. doi:10.1029/ (2013) Radar interferometer calibration of the EISCAT Svalbard Radar and a JA083iA09p04131 additional receiver station. J Atmos Sol Terr Phy 105–106:287–292 Ridley AJ, Deng Y, Toth G (2006) The Global Ionosphere-Thermosphere Model Schmidt H, Brasseur GP, Charron M, Manzini E, Giorgetta MA, Diehl T, Fomichev (GITM). J Atmos Solar-Terrestr Phys 68:839–864 VI, Kinnison D, Marsh D, Walters S (2006) The HAMMONIA chemistry climate Rietveld MT, Goncharov NP (1998) Artificial periodic irregularities from the model: sensitivity of the mesopause region to the 11-year solar cycle and Tromso heating facility. Adv Space Res 21(5):693–696 CO2 doubling. J Clim 19:3903–3931 Rietveld MT, Kopka H, Stubbe P (1986) D-region characteristics deduced from Schoeberl MR, Lait LR, Newman PA, Rosenfield JE (1992) The structure of the pulsed ionospheric heating under auroral electrojet conditions. J Atmos Terr polar vortex. J Geophys Res 97:7859–7882 Phys 45:311–326 Schumann MR, Roebber PJ (2010) The influence of upper-tropospheric potential Rietveld MT, Kosch MJ, Blagoveshchenskaya NF, Kornienko VA, Leyser TB, vorticity on convective morphology. Mon Weather Rev 138:463–474. Yeoman TK (2003) Ionospheric electron heating, optical emissions, doi:10.1175/2009MWR3091.1,2010 and striations induced by powerful HF radio waves at high latitudes: Schunk RW et al. (2004) Global Assimilation of Ionospheric Measurements (GAIM). aspect angle dependence. J Geophys Res 108(A4):1141. Radio Sci 39(1):Artn Rs1s02. doi:10.1029/2002rs002794 doi:10.1029/2002JA009543 Schunk RW, Scherliess L, Thompson DC (2011) Ionosphere data assimilation: Rishbeth H, Roble RG (1992) Cooling of the upper atmosphere by enhanced problems associated with missing physics, aeronomy of the Earth’s greenhouse gases: modelling of thermospheric and ionospheric effects. atmosphere and ionosphere, part 5. IAGA Special Sopron Book Series Planet Space Sci 40:1011–1026 2:437–442. doi:10.1007/978-94-007-0326-1_3 Robinson TR, Strangeway R, Wright DM, Davies JA, Horne RB, Yeoman TK, Seppälä A, Verronen PT, Kyrölä E, Hassinen S, Backman L, Hauchecorne A, Bertaux Stocker AJ, Lester M, Rietveld MT, Mann IR, Carlson CW, McFadden JP (2000) JL (2004) Solar proton events of October–November 2003: ozone depletion FAST observations of ULF waves injected into the magnetosphere by means in the Northern Hemisphere polar winter as seen by GOMOS/Envisat. of modulated RF heating of the auroral electrojet. Geophys Res Lett Geophys Res Lett 31, L19107. doi:10.1029/2004GL021042 27(19):3165–3168 Seppälä A, Verronen PT, Sofieva VF, Tamminen J, Kyrölä E, Rodger CJ, Clilverd MA Roble RG (2000) On the feasibility of developing a global atmospheric model (2006) Destruction of the tertiary ozone maximum during a solar proton event. extending from the ground to the exosphere. Atmospheric science across Geophys Res Lett 33:L07804. doi:10.1029/2005GL025571 the stratopause. Geophys Monogr Amer Geophys Union 123:53–68 Seppälä A, Randall CE, Clilverd MA, Rozanov E, Rodger CJ (2009) Geomagnetic Roble RG, Dickinson RE (1989) How will changes in carbon dioxide and methane activity and polar surface air temperature variability. J Geophys Res D modify the mean structure of the mesosphere and thermosphere? Geophys 114:A10312. doi:10.1029/2008JA014029 Res Lett 16:1441–1444 Shea MA, Smart DF, McCracken KG, Dreschhoff GAM, Spence HE (2006) Solar Roble RG and Tzur I (1986) The global atmospheric-electrical circuit. In: The proton events for 450 years: the Carrington event in perspective. Adv Space Earth’s electrical environment. National Research Council, Geophysics Res 38(2):232–238. doi:10.1016/j.asr.2005.02.100, ISSN 0273–1177 Study Committee, American Geophysical Union Shepherd TG, Koshyk JN, Ngan K (2000) On the nature of large-scale mixing in Rodger AS, Moffet RM, Quegan S (1992) The role of ion drift in the formation of the stratosphere and mesosphere. J Geophys Res 105(D10):12433–12446. ionisation troughs in the mid- and high-latitude ionosphere: a review. J doi:10.1029/2000JD900133 Atmos Terr Phys 54:1–30 Shindell DT, Rind D, Lonergan P (1998) Increased polar stratospheric ozone losses Rossby CG et al (1939) Relations between variations in the intensity of the zonal and delayed eventual recovery owing to increasing greenhouse-gas circulation of the atmosphere and the displacements of the semipermanent concentrations. Nature 392:589–592 centers of action. J Marine Res 2:38–55 Shine KP (1986) On the modelled thermal response of the Antarctic stratosphere Rossi A, Anselmo L, Cordelli A, Farinella P, Pardini C (1998) Modelling the evolution to a depletion of ozone. Geophys Res Lett 13:1331–1334 of the space debris population. Planet Space Sci 46(11–12):1583–1596 Simmons AJ (1974) Planetary-scale disturbances in the polar winter stratosphere. Röttger J (2001) About multiple layering and aspect sensitivity of polar Quart J Roy Met Soc 100:76–108 mesosphere summer echoes. Proc. MST10 meeting, Jicamarca, Singer W, Bremer J, Hocking WK, Weiss J, Latteck R, Zechal M (2003) Temperature http://jro.igp.gob.pe/mst10/CD/ExtAbs/Session1/I1_009.pdf and wind tides around the summer mesopause at middle and arctic Rozanov E, Callis L, Schlesinger M, Yang F, Andronova N, Zubov N (2005) latitudes. Adv Space Res 31(9):2055–2060 Atmospheric response to NOy source due to energetic electron precipitation. SmirnovaM,BelovaE,KirkwoodS,Mitchell N (2010) Polar mesosphere Geophys Res Lett 32:L14811. doi:10.1029/2005GL023041 summer echoes with ESRAD, Kiruna, Sweden: Variations and trends over McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 62 of 63

1997-2008. J Sol Atmos Terry Phys 72:435-447. doi:10.1016/ Tulasi Ram S, Lei J, Su SY, Liu CH, Lin CH, Chen WS (2010) Dayside ionospheric j.jastp2009.12.014 response to recurrent geomagnetic activity during the extreme solar Smirnova M, Belova E, Kirkwood S (2011) Polar mesosphere summer echo minimum of 2008. Geophys Res Lett 37:L02101. doi:10.1029/2009GL041038 strength in relation to solar variability and geomagnetic activity during Turunen T (1986) GEN-SYSTEM—a new experimental philosophy for EISCAT 1997–2009. Ann Geophys 29:563–572 radars. J atmos terr Phys 48:777–785 Smirnova M, Belova E, Kirkwood S (2012) Aspect sensitivity of polar mesosphere Turunen E, Matveinen H, Tolvanen J, Ranta H (1996) D-region ion chemistry summer echoes based on ESRAD MST radar measurements in Kiruna. Ann model. In: Schunk W (ed) STEP handbook of ionospheric models., pp 1–25 Geophys 30:457–465 Ulich T, Turunen E (1997) Evidence for long-term cooling of the upper atmosphere Sojka JJ, Raitt WJ, Schunk RW (1981) A theoretical study of the high-latitude in ionospheric data. Geophys Res Lett 24:103 winter F region at solar minimum for low magnetic activity. J Geophys Valladares CE, Basu S, Buchau J, Friis-Christensen E (1994) Experimental evidence for Res 86:609–621 the formation and entry of patches into the polar cap. Radio Sci 29:167–194 Solomon S, Rusch DW, Gerard JC, Reid GC, Crutzen PJ (1981) The effect of Vallinkoski M, Lehtinen MS (1991) Behaviour of parameter estimation errors particle precipitation events on the neutral and ion chemistry of the middle in linear statistical inversion theory with applications to incoherent atmosphere: II. Odd hydrogen. Planet Space Sci 8:885–893 scatter error estimates. J Atmos Terr Phys 52:675–685 Solomon S, Reid GC, Rusch DW, Thomas RJ (1983) Mesospheric ozone depletion Vanhamäki H, Amm O, Viljanen A (2007) Role of inductive electric fields and currents during the solar proton event of July 13, 1982: 2. Comparisons between in dynamical ionospheric situations. Ann Geophys 25:437–455 theory and measurements. Geophys Res Lett 10:257–260 Vassiliades D (2000) System identification, modelling and prediction for Solomon SC, Woods TN, Didkovsky LV, Emmert JT, Qian L (2010) space weather environments. IEEE T Plasma Sci 28(6):1944–1955 Anomalously low solar extreme-ultraviolet irradiance and thermospheric Verronen PT, Turunen E, Ulich T, Kyrölä E (2002) Modelling the effects of the density during solar minimum. Geophys Res Lett 37:2010. doi:10.1029/ October 1989 solar proton event on mesospheric odd nitrogen using a 2010GL044468 detailed ion and neutral chemistry model. Ann Geophys 20:1967–1976 Song Y, Robinson WA (2004) Dynamical mechanisms for stratospheric Verronen PT, Seppälä A, Kyrölä E, Tamminen J, Pickett HM, Turunen E (2006) influences on the troposphere. J Atmos Sci 61:1711–1725. doi:10.1175/ Production of odd hydrogen in the mesosphere during the January 2005 solar 1520 0469(2004)061 proton event. Geophys Res Lett 33:L24811. doi:10.1029/2006GL028115 Spogli L, Alfonsi L, De Franceschi G, Romano V, Aquino MHO, Dodson A (2009) Vierinen J and Lehtinen MS (2009) 32-cm wavelength radar mapping of the Climatology of GPS ionospheric scintillations over high and mid-latitude moon. Proceedings of the 6th European Radar Conference – European regions. Ann Geophys 27:3429 3437. doi:10.5194/angeo-27-3429 Vierinen J, Lehtinen MS, Markkanen J, Virtanen II (2009a) Measuring space Stanislawska I, Belehaki A (2009) Space weather observational activities and data debris with phase coded aperiodic transmission sequences. Proc. 5th management in Europe. Acta Geophysica 57:1 ESA space debris conference Stebel K, Blum U, Fricke KH, Kirkwood S, Mitchell NJ, Osepian A (2004) Joint Vierinen J, Markkanen J, Krag H (2009b) High power large aperture radar observations radar/lidar observations of possible aerosol layers in the winter mesosphere. of the Iridium-COSMOS collision, Proc. 5th ESA space debris conference – J Atmos Solar Terr Phys 66:957 970 Vierinen J, Virtanen II, Lehtinen MS, Roininen L, Raita T, Amm O, van de Stevens MH, Gumbel J, Englert CR, Grossmann KU, Rapp M, Hartogh P (2003) Kamp M, Kauristie K and Luntama JP (2010) TOMOSCAND—amesoscale, Polar mesospheric clouds formed from space shuttle exhaust. Geophys Res multi-frequency 3D ionospheric tomography network for Fenno-Scandinavia. Lett 30(10):1546. doi:10.1029/2003GL017249 Proceedigs of the International Beacon Satellite Symposium, BSS2010, P. Strelnikov B, Rapp M, Zecha M, Blix TA, Friedrich M, Yeoman TK (2009) PMSE and Doherty, M. Hernandez-Pajares, J.M. Juan, J. Sanz and A. Aragon-Angel (eds), E-region plasma instability: in situ observations. J Atmos Sol Terr Phys Campus Nord UPC, Barcelona, 2010 71(1):143–157 Svensmark H (1998) Influence of cosmic rays on Earth’s climate. Phys Rev Viljanen A (2011) European project to improve models of geomagnetically Lett 81:5027–5030 induced currents. Space Weather 9:S07007. doi:10.1029/2011SW000680 Svensmark H, Friis-Christensen E (1997) Variation of cosmic ray flux and global cloud Viljanen A, Tanskanen EI (2011) Climatology of rapid geomagnetic variations – coverage—a missing link in solar-climate relationships. Journal of Atmos Sol Terr at high latitudes over two solar cycles. Ann Geophys 9:1783 1792. Phys 59(11):1225–1232 doi:10.5194/angeo-29-1783-2011 Szasz C (2008) Radio meteors above the Arctic Circle: radiants, orbits and Viljanen A, Amm O, Pirjola R (1999) Modelling geomagnetically induced currents – estimated magnitudes. Thesis, IRF Scientific Report 294 during different ionospheric situations. J Geophys Res 104:28059 28071 Szasz C, Kero J, Meisel DD, Pellinen-Wannberg A, Wannberg G, Westman A (2008) Viljanen A, Nevanlinna H, Pajunpää K, Pulkkinen A (2001) Time derivative of Orbit characteristics of the tristatic EISCAT UHF meteors. Monthly Notices of the horizontal geomagnetic field as an activity indicator. Ann Geophys – Royal Astronomical Society 388(1):15–25 19:1107 1118 Tamburini F, Mari E, Sponselli A, Thide B, Bianchini A, Romanato F (2012) Viljanen A, Tanskanen EI, Pulkkinen A (2006) Relation between substorm Encoding many channels on the same frequency through radio characteristics and rapid temporal variations of the ground magnetic – vorticity: first experimental test. New J Phys 14:033001. doi:10.1088/ field. Ann Geophys 24:725 733 1367-2630/14/3/033001 Vincent RA (1984) MF/HF radar measurements of the dynamics of the — – Tanskanen E, Pulkkinen TI, Koskinen HEJ (2002) Substorm energy budget during mesopause region a review. J Atmos Terr Phys 46(11):961 974 low and high solar activity: 1997 and 1999 compared. J Geophys Res Virtanen II, Lehtinen MS, Nygrén T, Orispää M, Vierinen J (2008) Lag profile – 107:1086. doi:10.1029/2001JA900153 inversion method for EISCAT data analysis. Ann Geophys 26:571 581 Thomas GE (1996) Is the polar mesosphere the miner’s canary of global change? Virtanen II, Vierinen J, Lehtinen MS (2009) Phase-coded pulse aperiodic transmitter – Adv Space Res 18(3):149–158 coding. Ann Geophys 27:2799 2811 Thompson TW (1978) High resolution lunar radar map at 7.5 meter wavelength. Virtanen II, Lind FD, Roininen L, Erickson PJ, Rideout WC, Orispää M, Vierinen Icarus 36:174–188 J, Lehtinen MS (2013) Polyphase-coded incoherent scatter measurements Thrane EV (1986) DC-probe measurements in the middle atmosphere Middle at Millstone Hill. Radio Sci 48(5):519–526 Atmosphere Program, 6, vol 19. SCOSTEP, University of Illinois, Urbana, IL Voiculescu M, Virtanen II, Nygrén T (2006) The F-region trough: seasonal morphology Tinsley BA (2008) The global atmospheric electric circuit and its effects on cloud and relation to interplanetary magnetic field. Ann Geophys 24:173–185 microphysics. Rep Prog Phys 71(6):066801 Voiculescu M, Nygrén T, Aikio AT, Kuula R (2010) An olden but golden Tinsley BA, Heelis RA (1993) Correlations of atmospheric dynamics with solar EISCAT observation of a quiet‐time ionospheric trough. J Geophys Res activity evidence for a connection via the solar wind, atmospheric 115:A10315. doi:10.1029/2010JA015557 electricity, and cloud microphysics. J Geophys Res 98:10375–10384. WälchliU,StegmanJ,WittG,ChoJYN,MillerCA,KelleyMC,SwartzWE doi:10.1029/93JD00627 (1993) First height comparison of noctilucent clouds and simultaneous Tsurutani BT, Mannucci AJ, Iijima B, Guarnieri FL, Gonzalez WD, Judge DL, PMSE. Geophys Res Lett 20(24):2845–2848. doi:10.1029/93GL02631 Gangopadhyay P, Pap J (2006) The extreme Halloween 2003 solar flares (and Walsh AP, Fazakerley AN, Lahiff AD, Volwerk M, Grocott A, Dunlop MW, Lui Bastille Day, 2000 flare), ICMEs, and resultant extreme ionospheric effects: a AT, Kistler LM, Lester M, Mouikis C, Pu Z, Shen C, Shi J, Taylor MGGT, review. Adv Space Res 37:1583 Lucek E, Zhang TL, Dandouras I (2009) Cluster and McCrea et al. Progress in Earth and Planetary Science (2015) 2:21 Page 63 of 63

multipoint observations of a plasma bubble. Ann Geophys 27:725–743. doi:10.5194/angeo-27-725-2009 Wannberg G (1993) The G2-system and general purpose alternating code experiments for EISCAT. J atmos terr Phys 55:543–557 Wannberg G, Pellinen-Wannberg A, Westman A (1996) An ambiguity-function based method for analysis of Doppler decompressed radar signals applied to EISCAT measurements of oblique UHF-VHF meteor echoes. Radio Sci 31:497–518 Wannberg UG, Andersson H, Behlke R, Belyey V, Bergqvist P, Borg J, Brekke A, Delsing J, Eliasson L, Finch I, Grydeland T, Gustavsson B, Haggstrom I, Harrison RA, Iinatti T, Johansson G, Johansson J, Johansson J, La Hoz C, Laakso T, Larsen R, Larsmark M, Lindgren T, Lundberg M, Markkanen J, Marttala I, McCrea IW, McKay D, Postila M, Puccio W, Renkwitz T, Turunen E, van Eyken A, Vanhainen LG, Westman A, Wolf I (2010) EISCAT_3D—a next-generation European radar system for upper atmosphere and geospace research. Radio Science Bulletin 332:75–88 Webb DF, Allen JH (2004) Spacecraft and ground anomalies related to the October-November 2003 solar activity. Space Weather 2(3). doi:10.1029/ 2004SW000075 Wehrenpfennig A, Jakowski N, Wickert J (2001) A dynamically configurable system for operational processing of space weather data, physics and chemistry of the earth, part C: solar. Terr Planet Sci 26(8):601–604 Weimer DR (2005) Improved ionospheric electrodynamic models and application to calculating Joule heating rates. J Geophys Res 110:A05306. doi:10.1029/ 2004JA010884 Whalen JA (1989) The daytime F layer trough and its relation to ionospheric-magnetospheric convection. J Geophys Res 94:17169–17184 Wilink TJ, Davies NC, Angling MJ, Jodalen V, Lundborg B (1999) Robust HF data communications at high latitudes. IEE Proc-Microw Antennas Propag 146:263–268 Woodfield EE, Wild JA, Kavanagh AJ, Senior A, Milan SE (2010) Combining incoherent scatter radar data and IRI‐2007 to monitor the open‐closed field line boundary during substorms. J Geophys Res 115:A00I15. doi:10.1029/ 2010JA015751 Wrixon AD (2008) New ICRP recommendations. J Radiol Prot 28:161–168. doi:10.1088/0952-4746/28/2/R02 Yau AW, André M (1997) Sources of ion outflow in the high latitude ionosphere. Space Sci Rev 80:1–25. doi:10.1023/A:1004947203046 Yau AW, Abe T, Peterson WK (2007) The polar wind: recent observations. J Atmos Sol Terr Phys 69:1936–1983 Yeh KC, Liu CH (1982) Radio wave scintillations in the ionosphere. Proc IEEE 70(4):324–360 Yu TY, Palmer RD, Chilson PB (2001) An investigation of scattering mechanisms and dynamics in PMSE using coherent radar imaging. J Atmos Terr Phys 63(17):1797–1810 Yue J, Liu HL (2010) Fast meridional transport in the lower thermosphere by planetary-scale waves. J Atmos Sol-Terr Phys 72:1372–1378 Zank G (2012) The space radiation environment in energetic particles at the earth. Bull Am Phys Soc 2012(57):1 Zeller O, Zecha M, Bremer J, Latteck R, Singer W (2006) Mean characteristics of mesosphere winter echoes at mid- and high-latitudes. J Atmos Solar Terr Phys 68:1087–1104 Zhang S-R, Holt JM (2013) Long-term ionospheric cooling: dependency on local time, season, solar activity, and geomagnetic activity. J Geophys Res 118:3719–3730. doi:10.1002/jgra.50306 Zolesi B, Cander LR (2004) COST 271 action effects of the upper atmosphere on terrestrial and Earth-space communications. Radio Sci 39:Rs2011 Submit your manuscript to a journal and benefi t from:

7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the fi eld 7 Retaining the copyright to your article

Submit your next manuscript at 7 springeropen.com